In a 5-l. round-bottomed flask are placed 2 kg. (15.9 moles) of pure crystallized oxalic acid which has been passed through a 10-mesh sieve, and 3 l. of carbon tetrachloride (Note 1). The flask is then fitted (Note 2) with a mechanical stirrer and the apparatus shown in Fig. 22. This apparatus consists of an upright water condenser through the jacket of which steam is passed. A connecting tube leads from the top of this condenser to the top of a second efficient spiral condenser (Note 3), the bottom of which extends into a gravity separator. The water runs off into a receiver, and the heavier carbon tetrachloride layer returns to the reaction flask. A few essential points in the apparatus should be mentioned: it is necessary that the return arm of the gravity separator be below the tube through which the water flows into the receiver; the bottom of the funnel should be below the water outlet as in the diagram; the openings A and B are necessary to prevent siphoning in the respective tubes; and the last bend in the return tube before entering the reaction flask serves as a trap to prevent vapors from entering that tube.

The flask is placed up to its neck (Note 4) in hot water, stirring is started, and the steam is allowed to flow through the first upright condenser. The heating of the flask (Note 5) and vigorous stirring are continued until no more water passes through into the condenser. This point is indicated by the fact that no more water flows from the side arm of the separator. A slight milkiness in the carbon tetrachloride, which persists even when practically no more water comes over, need not be considered. The time of heating depends to a large extent on the rate with which the vapors are distilled over, and this rate in turn depends largely on the efficiency of the condenser. With an ordinary apparatus, such as that described, the time necessary will be approximately eighteen to twenty-four hours.

The reaction mixture is now filtered with suction, care being taken that any oxalic acid which is crusted above the surface of the carbon tetrachloride is not removed with the main portion of the product. The carbon tetrachloride may be used again for another run (Note 6), and the anhydrous oxalic acid is dried by short exposure to a warm dry atmosphere. The yield of product amounts to more than 90 per cent of the theoretical amount.

A laboratory oven (Note 1) is equipped with as many clay plates or enameled pie plates or trays as it will accommodate and is adjusted to operate at 98–99° (Note 2) and (Note 3). When the temperature has become constant the plates are removed, rapidly covered with a layer (not over 3–4 mm. deep) of pulverized (Note 4) hydrated oxalic acid, and then quickly replaced in the oven. The temperature will drop slightly for a few minutes (Note 5). After the oven has regained the temperature for which it was adjusted, it is heated for two hours longer at this temperature. The product is then removed, crushed if slightly caked, and quickly bottled. The yield from 100 g. of hydrated oxalic acid is 69–70 g. (96–98 per cent of the theoretical amount) (Note 6). The product is 99.5–100 per cent pure, as indicated by titration with standard alkali (Note 7).