Galileo (1564-1642) first taught that the air has weight. His pupil Torricelli went a step further. Torricelli saw that the column of water, held up by the pressure of the air in the tube of the pump, must exactly balance a similar column of air, reaching from the surface of the water in the well to the top of the atmosphere. The column of water, in other words, exactly replaces this column of air. While working on this subject, Torricelli, in 1643, performed the following experiment. He filled a glass tube, about 3 feet long and closed at one end, with mercury. After filling the tube, he put his finger over the open end and inverted the tube over a vessel containing mercury. When the lower end of the tube was below the surface of the mercury in the dish, he removed his finger. At once the column of mercury fell in the tube until it stood at a height of about 30 inches, leaving a vacant space of 6 inches in the upper part of the tube. This space has since been known as the _Torricellian vacuum_. Torricelli had proved what he had expected, viz., that the height of the column of liquid which replaces and balances an air column of the same size varies with the weight of that liquid. It takes a column of water 34 feet long to balance a similar column of air. It takes a column of mercury only 30 inches long to balance a similar column of air. This, as Torricelli correctly explained, is due to the fact that mercury is so much (13-1/2 times) heavier than water. The column of water weighs just the same as the column of mercury. Each column exactly balances an air column of similar cross-section. The height of the water or of the mercury is a measure of the weight or pressure of the air. The greater the pressure on the surface of the water in the well, the higher will be the top of the water in the pump. The greater the pressure on the surface of the mercury in the basin, in the experiment of Torricelli, the higher will the mercury column stand in the glass tube. Either water or mercury may be used as the liquid in the barometer. Otto von Guericke (1602-1686), of Magdeburg, constructed a water barometer about 36 feet long, which he attached to the outside wall of his house. This barometer he used for some months, and made predictions of coming weather changes by means of it. A water barometer is, however, a very unwieldy thing to manage, on account of the great length of its tube. Furthermore, water barometers cannot be used in any countries where the temperatures fall to freezing. Mercury is the liquid universally employed in barometers. It is so heavy that only a small column of it is necessary to balance the atmospheric pressure. Therefore a mercurial barometer is portable. Further, mercury does not freeze until the temperature falls to 40° below zero.