In order practically to carry out this method of estimating carbonic acid the following apparatus is required:——Several bottles capable of holding 2·210 c. c., and well stoppered (failing bottles of exactly the right capacity Winchester quart bottles will answer); a small pair of bellows; several colourless glass cylinders marked at 100 c. c. capacity——the Nesslerising cylinders will answer for this purpose——a graduated pipette or burette to deliver tenths of a c. c. of solution, the standard solution of carbonate of soda, and the baryta water, which may be of moderate strength. The testing is managed thus: Winchester quart bottles having been made clean are rinsed with distilled water, and allowed to drain a little. They are then closed with their stoppers, and are ready for use. The operator having provided himself with two or three of these bottles and a small pair of bellows enters the room the air of which is to be tested. The stopper is then removed from one of the bottles, and some air of the room blown through with the bellows, and then the stopper is replaced, and the bottle carried away to be tested. The testing is done by pouring into the bottle 100 c. c. of clear baryta water, shaking up for two or three minutes, and then pouring out into a cylinder of colourless glass, and observing the depth of the turbidity in various lights and against various backgrounds. The turbidity is to be exactly imitated by means of the standard solution of carbonate of soda. In order to imitate the turbidity produced by a Winchester quart full of good air only 1 c. c. of this solution of carbonate of soda is required. If 2 c. c. or more than 2 are required, the air is bad and the ventilation is defective. In place of the first c. c. of solution of carbonate of soda the carbonic acid naturally present in a Winchester quart of good average air may be used, and a little practice and intelligence will suggest the necessary precautions.