By turning the valve, k, the vessel, b, could be partially filled with liquid oxygen, which, under a pressure of 10 mm. of mercury, boiled at about -210° C. Almost immediately the gaseous air began to condense and collect in the tube, a; a supply of fresh air was constantly maintained through the drying tubes, u and u', which were filled with sulphuric acid and soda lime respectively. When the quantity of liquid air ceased to increase, the tap on the U tube, u, was closed, the T tube, p' v', was connected with the manometer and air pump, and the liquid air was made to boil under a pressure of 10 mm. of mercury. In order to protect the liquid air from its warmer surroundings, a very thin, double wall tube, f, reaching to the level of the liquid in the outer vessel, was placed inside the tube, a. When, as in some of my experiments, liquid oxygen was used in the inner vessel, this part of the apparatus was dispensed with. [Illustration: Helium cooling apparatus] Using the apparatus I have just described, I carried out two series of experiments, in which liquid air and liquid oxygen were employed as cooling agents. The tube of the Cailletet's apparatus was thoroughly exhausted by means of a mercury pump, and then carefully filled with dry helium. In the first series of experiments the helium, confined under a pressure of 125 atmospheres, was cooled to the temperature of oxygen boiling, first under atmospheric pressure (-182.5°), and then under a pressure of 10 mm. of mercury (-210°). The helium did not condense under these conditions, and even when, as in subsequent experiments, I expanded the gas till the pressure fell to twenty atmospheres, and in some cases to one atmosphere, I could not detect the slightest indication that liquefaction had taken place. The first time that I compressed the gas I had, indeed, noticed that a small quantity of a white substance separated out and remained at the bottom of the tube when the pressure was released. Possibly this may have been due to the presence of a small trace of impurity in the helium, but it could not have constituted more than 1 per cent. of the total volume of the gas.