ured to a precision of one to two parts per 1000. Broadly speaking, this requires that the volumes be not less than 25 c.c. This factor of size of volume, however, cannot be considered the sole controlling factor of precision in titration methods as there are other considerations which enter into the problem. For a discussion of all these the reader is referred to Chapter VI.

37. Solution. In most analyses the next step after that of weighing out the appropriate amount of sample is that of obtaining its solution. Very often a knowledge of the nature of the substance will indicate at once the necessary procedure, but if such knowledge is lacking the general scheme is to try the solubility of the sample in water. If any insoluble portion remains, decant off the solution and treat the residue with acid; if any residue remains after the acid treatment, filter off the residue, wash it free of acid, ignite and then fuse it with an appropriate flux.

Treatment with acid. For substances other than metals the first acid to try is hydrochloric. If the sample contains reducing substances, the hydrochloric acid should be supplemented by nitric acid or some other oxidizing agent, such as potassium chlorate, bromine water, etc. On the other hand, if the sample contains oxidizing agents, the hydrochloric acid should be supplemented by some reducing agent, preferably sodium bisulphite added in small portions from time to time.

For metals an oxidizing agent is necessary in order to convert the metal from the elemental to the ionic form. It is usual to employ a mixture of nitric and sulphuric acids or to use sulphuric acid and add potassium chlorate in small portions during the process of solution. Certain non-ferrous alloys, particularly those containing tin, tungsten, or chromium, are often very resistant to the acid treatments just described. For such cases the use of hydrofluoric acid with nitric acid in the following manner is recommended. The weighed portion of the alloy is placed in a platinum dish of 100-125 c.c. capacity; next, 10-20 c.c. hydrofluoric acid are added, and then 0.5 c.c. or so of 16 M nitric acid at a time until about 5 c.c. have been added. Often the reaction is very vigorous upon the addition of the nitric acid so that caution must be used in adding this reagent. The efficiency of this method of effecting solution depends upon the fact that

fluoride ion forms highly undissociated or complex fluorides with so many of the metallic ions.2 The method carries the disadvantage, however, that the hydrofluoric acid must be removed after solution of the alloy is effected. To accomplish this removal a few c.c. of 18 M sulphuric acid are added to the solution, and the solution evaporated on the hot-plate until fumes of sulphur trioxide are just about ready to come off. (A few milligrams of platinum are always dissolved during this evaporation.)

The most effective concentration of acid to use in dissolving substances, either with respect to hydrochloric, nitric or sulphuric acid, is about 9 M; the amount of acid to use is about 2-3 c.c. of 9 M acid per 0.1 g. of sample.

38. Fusion. Substances which cannot be brought into solution by acid treatment must be submitted to fusion. The type of fusion which is employed depends largely on the nature of the substance and the further purpose of the analysis. There are four kinds of fusion ordinarily resorted to, namely:

The application of these fusions is given in the table on page 44, along with the particular flux to use, its relative amount, the kind of crucible necessary and the approximate temperature for the fusion.

In making any fusion there are certain points which should be observed. The substance should be not only in a fine state of subdivision (50 mesh or finer) but should be intimately mixed with the flux by grinding in an agate mortar. The mixture of the substance and the flux should not be put directly into the crucible, but the latter should first receive on its bottom and up its side walls as far as practicable a thin lining of the flux alone, as this often reduces the harmful effect of the fusion upon the crucible. The crucible should be large enough so that it is never more than one-half full and should always be kept covered. At the start of the heating a small flame should be used, and the temperature gradually raised until the necessary temperature has been attained. It is not advisable to heat to a much higher

2 Cf. Peters, Z. physik. Chem. 26, 193 (1898).

temperature than necessary because of the increasing deleterious effect upon the crucible. After the fusion, which usually takes about sixty minutes, has been effected, the molten mass should not be allowed to solidify as a solid cake in the bottom of the crucible but should be distributed in the form of a thin layer

around the walls of the crucible by taking the latter in a pair of tongs while the mass is still molten, withdrawing it from the flame and imparting a rotary and tilting motion to it, thereby causing the molten mass to flow up and around the walls of the crucible and to solidify there as a thin layer. This procedure greatly accelerates the subsequent solution of the fused mass. When cool, the crucible is placed on its side in a casserole and covered with water; acid is added or not as the case may be, the casserole covered with a watch-glass, and the temperature raised to 95°-100° and maintained until solution is accomplished.

39. Evaporation. The vessels used for evaporation of liquids are glass beakers, porcelain casseroles, and platinum dishes. It must be borne in mind that all kinds of glass are more or less soluble in water and in various solutions, the solubility increasing rapidly with the alkalinity of the solution and with increase of temperature, so that the use of glass beakers should be avoided. for evaporations which are preparatory to a precipitation. For such cases porcelain casseroles are far preferable to glass beakers owing to their lesser solubility, but where the work demands the highest precision, only platinum should be used. In no event should strongly alkaline solutions be evaporated as even platinum will be attacked under these conditions. Such solutions should first be neutralized or made faintly acid. Solutions faintly alkaline with ammonium hydroxide may be evaporated in porcelain without any sensible solution of the porcelain. In cases where it is necessary to evaporate a solution to dryness, the use of glass is always risky owing to the likelihood of its cracking when the solution goes dry. Porcelain is much safer to use.

During the evaporation of a liquid it should be protected from anything that might fall into it and, at the same time, the escape of the vapor must not be interfered with. These two desiderata can be accomplished by taking three or four semi-elliptic U-shaped loops of glass rod and hanging them equidistantly around the rim or edge of the evaporating dish and then placing on top of them a watch-glass of somewhat larger diameter than the dish.

See P. N. Walker, J. A. C. S. 27, 865 (1905), "Chemical Glassware," for composition, solubility and other behavior of chemical glassware.

'The limbs of the loops should be about 1-1.5 cm. long, have a diameter of about 0.6 cm., and a free distance between them of about 0.6 cm.

Solutions heated over a live flame are prone to bump so that evaporations should generally be conducted on a water-bath or a steam or gas hot-plate. Evaporations are at best time-consuming operations, frequently taking as long as eighteen hours; about the only thing the analyst can do to compensate matters is to employ himself with other operations during this time or else arrange matters so that the evaporations take place over night. Creeping. Certain salt solutions, particularly alkali chlorides, exhibit the phenomenon known as "creeping"; the salt deposits on the wall of the vessel at the surface of the solution and then gradually creeps up the wall and finally over the top of the vessel. This creeping can be stopped by painting a narrow band of collodion around the wall of the vessel a slight distance above the level of the solution.

40. Evaporation of Sulphuric or Hydrofluoric Acids. It is often required to treat a residue or precipitate in a crucible with concentrated sulphuric acid or with a few drops of sulphuric acid and 10-15 c.c. hydrofluoric acid, and then to get rid of the acids by evaporation. This operation is best accomplished by placing the crucible in an air bath and then maintaining a temperature of about 300°. A very serviceable and suitable air bath for this purpose may be made by taking a silica beakers about 7 cm. diameter and 9 cm. deep, and encompassing its upper edge by means of a triangle made of nichrome wire, this device enabling the operator to suspend the beaker from an iron ring. A quartz or platinum triangle with its legs bent at right angles to the plane of the triangle and cut to a length of about 3 cm. is placed inside of the beaker to receive the crucible. If the full flame of a Bunsen burner is used, a temperature of 350° may be attained inside the beaker. It will be found that an air bath of this description allows a very rapid evaporation without bumping or spattering, occurrences which almost invariably happen if one tries to conduct the evaporations by simply placing the crucible on an asbestos mat which is directly heated by the gas flame.

41. Precipitation. The general theory underlying precipitation will be found under the Solubility Product Principle so that we will confine ourselves to certain practical points of technique.

5 A beaker made out of monel metal answers equally well.