work. White says that neither the explosion method nor any other involving active combustion can give strictly accurate results. Jones and Parker,28 using the phenyl-disulphonic acid method whereby quantities of oxides of nitrogen as low as 0.001 c. c. could be detected, found that the production of oxides of nitrogen by the slow-combustion method when the time of burning was not more than 3 minutes and the wire not heated beyond a bright yellow, was within. the experimental error of routine gas analysis. Under the above conditions not more than 0.003 c. c. of oxides of nitrogen was produced. The purity of the platinum wire had no effect on the production of oxides of nitrogen. None was produced by the explosion method when air was used as the oxygen supply, but when mixtures of air and oxygen were used as the oxygen supply, appreciable quantities of oxides of nitrogen were formed, too large to be disregarded in gas analysis. The production of oxides of nitrogen by the explosion method is due to the flame temperature developed by the gas mixture when exploded rather than to the sparking across the electrodes. The addition of oxygen raises the flame temperature, which reaches a critical point above which appreciable quantities of oxides of nitrogen are produced.

DISTILLATION OF MERCURY

Apparatus for the distillation of mercury, devised by G. A. Hulett,20 former chief chemist of the Bureau of Mines, are used by the authors. His description of the apparatus is given below. The mercury is first filtered through chamois skin, then dropped in a fine stream through diluted nitric acid, and distilled in retorts heated by gas or electricity. Air is passed through the vapors to oxidize any volatile metals carried over in the distillate.

DESCRIPTION OF MERCURY STILL

Figure 19 gives an idea of the essential details of the gas-heated retort; g is a common asbestos air bath with a hole in the bottom for the flame; above is fastened a shallow metal cup that supports the flask and on which the flame plays. The flask a is of the ordinary round-bottom type holding 250 to 500 c. c. The neck is drawn down short, and the 20-cm. tube b sealed on, as is also the thin-walled side tube e, which is 10 to 15 mm. in diameter and 50 cm. long; e serves as the air-cooled condenser. A tube, c, is selected of such a diameter that it will snugly fit the tube b; to it is sealed a glass cock, and the

28 Jones, G. W., and Parker, W. L., The formation of oxides of nitrogen in the slowcombustion and explosion methods in gas analysis: Jour. Ind. Eng. Chem., vol. 13, 1921, p. 1154.

Hulett, G. A., The distillation of amalgams and the purification of mercury: Phys. Rev., vol. 34, 1911, p. 307.

tube is drawn out fine at the lower end. This tube is slipped into the tube b as indicated in the figure. The joint d, where there is a slight enlargement of the tube c, is made tight with rubber bands about 1 cm. wide, which are wrapped about the joint while stretched. The asbestos shields & deflect the hot gases so that this joint never gets warm, even to the hand. The joint is much simpler than a ground joint, which may be used.

The glass cock regulates the rate at which the air bubbles through the mercury in the still, and if it is properly ground no grease need be used on it. The end of the condensing tube e is sealed to the stopper of an ordinary Drexel washing bottle. The outlet tube from the Drexel bottle f is joined to an ordinary Sprengel suction

b

a

FIGURE 19.-Gas-heated mercury still. For explanation, see text

pump and a manometer, and a vacuum of 25 to 30 mm. is maintained in the system while air is bubbling through the mercury in the still. When the still is once uniformly in operation it needs little attention, and 2 kg. of mercury may be distilled in 2 or 3 hours. Steady gas and water pressures are desirable. The ground joints may readily be made tight enough by a little grinding with fine emery, if they are not satisfactory at the start. The rubber joint at d never gives trouble if the tube fits well in the neck of the flask.

DESCRIPTION OF STILL WITH ELECTRIC HEATER

In distilling large amounts of mercury it has been found necessary to use a larger still provided with an electric heater and so arranged

that mercury can be introduced during distillation. Figure 20 presents the main features of the still. The flask a is about 15 cm. in diameter, and into the bottom is sealed the glass tube q which connects with the reservoir h. The cock g is ordinarily left open. It was found that the air might be admitted from below through the side tube s instead of from above through the neck, as in the still of Figure 19. At f is a ground-glass joint. The inserted tube ends in a capillary which delivers the air well into the tube q, where it bubbles up through the mercury and into the still. With this

FIGURE 20.-Electrically heated mercury still. For explanation, see text

arrangement the rate at which the air is passing into the mercury can always be observed and controlled with the cock i. Mercury is allowed to flow into the reservoir h at any desired rate and thus the rate at which it enters the flask is controlled. The mercury vapor is condensed in the large U tube c, which is made from 25-mm. thin-walled tubing and is 50 cm. long, thus giving a condensing tube 1 meter long. The tube c is joined to a Sprengel suction pump and manometer, and a vacuum of 25 to 30 mm. is maintained in the apparatus during distillation. The mercury condenses in the tube c and collects in the tube d and flows out at the bent-up end. This tube d is about 1 meter long and 4 mm. in diameter.

ELECTRIC HEATER

The electric heater b is made of asbestos, water glass, and magnesium oxide cement, with heating coils of nichrome ribbon on the sides and bottom. The bottom part of the still is made from strips of asbestos 40 mm. wide and 1.5 mm. thick. The nichrome ribbon is placed along the edge of these strips, and then they are rolled up together, forming the ribbon into a spiral very close to one side of the "wheel" thus fabricated. The asbestos strips are well wet with water and the mixture of water glass and magnesium oxide, and after they are wound may, while still wet, be dished to fit the bottom of the flask. Thus the flask has a good support and the heating coil is at a place where it will be most effective. A short piece of brass tubing is attached to the end of the nichrome ribbon as a conveni

ence in starting the roll. This brass tube serves as a hole for the tube and also as one terminal for the heating coil. For 110-volt circuits this coil should have a resistance of 12 to 15 ohms.

A thin piece of asbestos is wound on a cylindrical form of the desired diameter. The asbestos wheel just made is forced into one end of the cylinder, and on its outside is wound nichrome ribbon to form the heater for the sides. A resistance of about 5 or 6 ohms is needed here. Around the cylinder is wound a layer of asbestos wet with the cement, and over all are wound several layers of asbestos for insulation. The top of the heater is also covered with asbestos and glass wool. More resistance will be

needed if all the heat from the 110-volt circuit is to be used in the heater, and a spiral of nichrome wire may be placed in the lower inside corner of heater k. The resistance of this coil may be adjusted so that no external resistance need be used. Normally the rate of distillation should be such that the mercury vapor is condensing over nearly the whole length of the condenser.

The apparatus is fastened to an upright support which is part of a large tray; a bracket holds the heater.

CARE OF APPARATUS

After the still is used, run out the last of the mercury and draw an acid up into the flask a, and the condensing tube if necessary (1:1 nitric acid is best for this purpose); after the acid, use distilled water. The still is easily cleaned and dried without dismantling. The glass cocks should be well ground to avoid the use of grease or organic matter, but the cocks need not be perfectly tight. Draw air through the mercury at the rate of about 5 c. c. per minute.

LABORATORY CONTAINER FOR HOLDING GASES OVER MERCURY

Figure 21 shows a container, which requires only a small amount of mercury, for holding gases that have been prepared in the laboratory. It follows the principle of one devised by Hempel,30 and is also similar in construction to Travers's,31 except that a three-way

30 Hempel, Walther, Methods of gas analysis. 3d German ed., trans. by L. M. Dennis, 1910, p. 29.

31 Travers, M. W., The experimental study of gases. 1901, p. 30.

stopcock has been sealed at the stop instead of the two-way cock which they used.

Remove cylindrical core from the block of hardwood d and cement a smaller cylindrical core, a, in place, as shown; the gas holder b

FIGURE 22.-Apparatus for determination of carbon monoxide by means of iodine pentoxide. For explanation, see text

fits closely over the inner cylinder. In using the instrument, first draw the mercury up into the stopcock. On account of the buoyancy of the glass exert pressure to expel the gas. First turn the threeway stopcock so that gas, as it is prepared, is passed through the