a platinum triangle or, lacking this, upon a silica or clay triangle; while hot it should be touched most preferably by platinum-tipped tongs, ess preferably by nickel tongs and never by iron.

Platinum crucibles should be frequently scoured both inside and out either with sea sand, Ottawa sand, or some abrasive of the general nature of a rouge. If this treatment fails to remove stains, particularly iron stains, the crucible may often be cleaned by a prolonged fusion of an hour or more with acid potassium sulphate. In this process the crucible is filled with the flux, the cover put on, and the temperature kept at a point (about 300°) where the fumes of SO, just begin to come off. Care must be taken, however, that the available SO, is not all driven off, leaving the normal sulphate, as this salt expands so rapidly on cooling that it might break the crucible. Sodium carbonate and borax may also be used as cleaning fluxes. For stains which resist the foregoing treatments the author has corroborated the experience of Dr. G. H. Walden, of this laboratory, that by leaving the crucible in an electric muffle furnace at a temperature around 1000° for 14 hours, the iron is brought to the surface, and then upon digestion with concentrated commercial hydrochloric acid is in most cases completely removed.

20. "Policeman." This is an instrument which is used in connection with precipitates for the purpose of detaching from the walls of the containing vessel any adhering particles which it is not possible to remove by the stream of water from a wash bottle. Essentially it is a stout piece of glass rod capped at one end with a piece of soft rubber tubing. The rod should be about 20 cm. long and should have a diameter between 5 and 7 mm.; the ends should be plane and annealed. The piece of rubber tubing should be about 2-3 cm. long and should have a bore small enough to insure a tight fit over the glass rod; the end of the tubing should project about 2-3 mm. beyond the rod, as this makes a flexible annular ring which adds greatly to the effectiveness of the "policeman." A policeman should never be used in place of a stirring rod to pour against nor should it be left standing in solutions.

21. Silver Nitrate Test Bottle. The silver nitrate test bottle is used for containing a solution of silver nitrate fairly acid with

nitric acid. The purpose of the solution is to test wash waters for evidence as to whether a precipitate has been completely washed or not. It is assumed in the washing of precipitates that the washing is completely accomplished and all soluble substances quantitatively removed when any particular substance originally present in the solution from which precipitation was made is quantitatively removed. Since in the great majority of cases precipitation is made from solutions containing chlorides, and the silver chloride test is a very delicate and satisfactory one, it follows that this test is the one that is most usually made.

The silver nitrate, nitric acid solution should be kept in a 100 c.c. ambercolored bottle. A cork stopper projecting about half its length above the top edge of the bottle should be employed and should be bored so as to accommodate a glass tube about 6 mm. diameter. This tube should extend nearly to the bottom of the bottle on the one hand and should project about 7 cm. above the top of cork on the other; its lower end should be drawn out somewhat so as to facilitate the formation of drops while the upper end should have its sharp edges rounded by holding in the gas flame for a while; a small plug of cotton wool should be placed in the upper end of the tube to prevent dust from getting into the bottle. The cork should not be jammed into the mouth of the bottle but should rest gently so that it can be lifted out with one hand.

22. Stirring Rods. These are slender glass rods which should be of assorted lengths with diameters about 4 mm. Both ends should have their sharp edges gently fused in the gas flame so that the rods will not cut the filter papers or scratch the beakers. The proper length for a stirring rod is such that the rod will not project more than 5 or 7 cm. beyond the lip of the beaker in which it is used when it rests diagonally on the floor of the beaker at a point directly opposite the lip. The student should provide himself with six to twelve stirring rods.

23. Wash Bottles. A wash bottle is a flat-bottomed flask of convenient size, fitted up to deliver a fine stream of distilled water or other washing medium for the purpose of washing precipitates. One wash bottle for use with water at ordinary room temperatures will always be needed, and it is advisable to have another for use when hot water (70°-90°) is needed; in addition

20.2 molar with respect to silver nitrate and 0.5 molar with respect to nitric acid, or 3.5 g silver nitrate and 3 c.c. concentrated nitric acid (sp. gr. 1.42) per 100 c.c.

to these two wash bottles for water it is well to provide several smaller wash bottles for use with liquids other than water. Flatbottomed flasks are invariably used, and the following sizes will be found convenient: for water a flask of about 750 c.c. capacity; for the other liquids one of 500 c.c. capacity. The quality of the glass is very important as many kinds of glass are appreciably soluble, particularly in hot water. Pyrex glass, however, is extremely insoluble in water and is admirably suited for wash bottles. The accompanying diagram shows the general arrangement of a wash bottle.

In the assemblage of a wash bottle the following points are to be observed: the rubber stopper (2-hole) when set firmly in place should project about half its length above the top edge of the flask; if it projects more than this it is likely to pop out when the wash bottle is being used, while if less it will be troublesome to remove. The mouthpiece should be bent at an angle of 45° and should just go through the stopper. The lower end of the tube should extend to the bottom of the flask and should be curved forward as shown because this design allows of more complete emptying of the contents of the wash bottle when in use; the upper end of the tube should be bent at an angle of 135° and in the same plane as the lower end; the top of the bend should be about 3 cm. above the top of the cork and the tube should extend about 4 cm. from the top of the bend. It is connected to the nozzle by means of a piece of soft rubber tubing. The nozzle should be about 4-5 cm. long and should be drawn out to a taper, having an orifice about 1 mm. diameter. This dimension can be arrived at by holding the tapered end in a small flame and contracting the orifice by brief incipient melting until such a size is reached as upon trial the nozzle will deliver about 10 c.c. water in 10 seconds when the full pressure of the breath is applied and maintained at the mouthpiece.

FIG. 2

WASH BOTTLE

For hot water bottles certain further points are to be observed. In order to protect the hand, the neck of the bottle should have fastened around it some suitable insulating material like sheet cork or asbestos paper, or a wrapping of heavy cord. In order to protect the mouth from scalding by

3

The asbestos paper can be easily and satisfactorily applied as follows: take a strip about 3 in. wide and 12-15 in. long and drench it with water; while it is still wet wrap it snugly around the neck of the bottle and compress the several layers together by pressure of the hand; let dry over night, and the bottle will be ready next morning. There is enough natural cement in asbestos paper to furnish its own binder.

the back rush of steam through the mouthpiece when the pressure of the breath is removed, a three-hole rubber stopper should be used and a short piece of glass tubing, open at both ends, inserted in the extra hole. By keeping the thumb over this piece of tubing while the water is being blown out, and then lifting the thumb before the blowing is stopped, there will be no back rush of steam into the mouth of the analyst. In the continuous use of a hot water bottle, the steam seems to disintegrate the rubber cork so that small pieces of the cork drop into the water and thus contaminate it or else find their way through the outgoing stream of water into the precipitate being washed; consequently, the condition of the cork should be frequently examined and a new cork used when necessary.

24. Reagents. A list of the reagents which are required for the ordinary purposes of a course in quantitative analysis is given at the end of the chapter; the quantities mentioned there are those which have been found to suffice for the average needs of a student for a laboratory course of 150 hours.

Purity of Reagents. In regard to this question it cannot be emphasized too strongly that an analyst should always take the greatest pains to make certain of the purity of his reagents. To this end he should always employ only the best grades of "analyzed chemicals." Even then he should not accept the published analysis on the label as his final authority, but should check matters up by running a blank test for any impurity, whether it is listed or not, that would interfere in the analysis to be performed.4

While the "analyzed chemicals" which are now obtainable for quantitative work mark a distinct advance over those merely labelled "C. P.," in that the impurities are not only stated on the label but their percentages given, it must always be borne in mind that a label is by no means an infallible criterion. There are several reasons for making this assertion. In the original

For the testing of chemical reagents, the student is referred to the admirable work of C. Krauch, Chemical Reagents, Their Uses, Methods of Testing for Purity and Commercial Varieties. Second English Edition, Revised and enlarged by H. B. Stocks. Scott Greenwood & Son, London, 1919, 369 pp.

For methods dealing with the preparation of pure reagents for analytical purposes there does not seem to be any single collected work although there are several books on preparation of inorganic salts, particularly L. Vanino, Handbuch der Präparativen Chemie, Band I, Anorganische Teil. Ferd. Enke, Stuttgart, 1921, 812 pp. Also N. and W. Biltz, Laboratory Methods of Inorganic Chemistry. Translated by W. T. Hall and A. A. Blanchard. John Wiley & Sons, New York, 1909, 258 pp. In the main it is mostly a question of searching the literature for articles dealing with the particular reagent in question.

unopened bottle certain impurities might not have been tested for by the manufacturer, while very often dust and stray pieces of straw or cork will be present in spite of the greatest care used in packing; in bottles which have been opened there is always the chance that some one may have left the bottle uncorked for a time with consequent contamination of the contents from the fumes of the laboratory, or else an unused portion of the reagent may have unwittingly been returned to the package or through gross mistake may even have been put in a wrong package; in packages which have been dispensed from others there are in addition to the foregoing chances of impurity the ones that a false label may have been applied or that the new package might not have been thoroughly clean. It must further be remembered that alkaline solutions kept for any length of time in glass containers will be contaminated with soluble constituents from the glass.

To show the degree of purity which may be expected in "analyzed chemicals" for quantitative use, we have listed at the end of the chapter certain typical analyses as supplied by the manufacturers for some of the more common reagents; those reagents which are marked with an asterisk are used as primary standards in analytical work.

There are several ways

25. Concentration of Reagents. in use for designating the concentration of a reagent, and these are discussed in detail in Chapter VIII. The system that we shall use now and throughout our work will be the molar system. The units of this scheme are the mole and the liter; a mole being the formula weight in grams of a substance, a molar solution being one which contains one mole of reagent per liter of solution; thus a molar solution of silver nitrate (mol. wt. 169.9) is one which contains 169.9 g. silver nitrate per liter of solution, a tenth-molar solution of hydrochloric acid (mol. wt. 36.46) is one which contains 3.646 g. HCl per liter of solution. The symbol M is used for molar, whence it follows that the concentration of a solution designated as 5 M, would contain 5 moles of a substance per liter of solution; and one marked, 0.2 M would indicate a concentration of two-tenths of a mole of substance per liter of solution. For general analytical purposes the following concentrations