Porcelain filter candles are examined with reference to their power of holding back _all_ the micro-organisms suspended in the fluids which are filtered through them, and permitting only the passage of germ-free filtrates. In order to determine the freedom of the filter from flaws and cracks which would permit the passage of bacteria no matter how perfect the general structure of the candle might be, the candle must first be attached by means of a long piece of pressure tubing, to a powerful pump, such as a foot bicycle pump, fitted with a manometer. The candle is then immersed in a jar of water and held completely submerged whilst the internal pressure is gradually raised to two atmospheres by the action of the pump. Any crack or flaw will at once become obvious by reason of the stream of air bubbles issuing from it. The examination for permeability is conducted as follows: _Apparatus Required_: Filtering apparatus: The actual filter candle that is used must be the one it is intended to test and must be previously carefully sterilised; the arrangement of the apparatus will naturally vary with each different form of filter, one or other of those already described (_vide_ pages 42-48). Plate-levelling stand. Case of sterile plates. Case of sterile pipettes, 10 c.c. (in tenths). Case of sterile pipettes, 1 c.c. (in tenths). Tubes of nutrient gelatine. Flask containing sterile normal saline solution. Sterile measuring flask, 1000 c.c. capacity. METHOD.-- 1. Prepare surface cultivations, on nutrient agar in a culture bottle, of the Bacillus mycoides, and incubate at 20° C., for forty-eight hours. 2. Pipette 5 c.c. sterile normal saline into the culture bottle and emulsify the entire surface growth in it. 3. Pipette the emulsion into the sterile measuring flask and dilute up to 1000 c.c. by the addition of sterile water. 4. Pour the emulsion into the filter reservoir and start the filtration. 5. When the filtration is completed, pour six agar plates each containing 1 c.c. of the filtrate. 6. Incubate at 37° C. until, if necessary, the completion of seven days. 7. If the filtrate is not sterile, subcultivate the organism passed and determine its identity with the test bacterium before rejecting the filter--since the filtrate may have been accidentally contaminated. 8. If the filtrate is sterile, resterilise the candle and repeat the test now substituting a cultivation of B. prodigiosus--a bacillus of smaller size. 9. If the second test is satisfactory, test the candle against a cultivation of a very small coccus, e. g., Micrococcus melitensis, in a similar manner; in this instance continuing the incubation of cultivations from the filtrate for fourteen days. TESTING OF DISINFECTANTS.