leading to the pump, protected by a brass cap and valve. The porosity of wood may be shown by the mercury funnel, in which mercury is driven lengthwise through a piece of wood which passes through the top of a receiver. If a piece of wood held under water is covered with a receiver, and the air exhausted, torrents of bubbles imprisoned in its pores will pour from it. Now on admitting the air the water enters the wood, which becomes water-logged, and no longer floats. The revolving jet is a bent brass tube, like a Barker's Mill, which when placed under a receiver turns rapidly in one direction when the air is exhausted, and in the other when it is readmitted. The effect of the resistance of the air is shown by two fan wheels with vanes set flatwise and edgewise, respectively. If set in motion, the former stops first in air, but both revolve for nearly the same time in a vacuum. The same effect may be shown by a feather and guinea placed in a long glass tube from which the air is removed. They then fall with nearly equal velocity from end to end. An important experiment is the proof of the weight of the air. A glass sphere is weighed when full of air and when exhausted, and the difference gives approximately the required weight. The exact weight is obtained only by an accurate correction for temperature, pressure and moisture. The two following experiments, though properly belonging to other branches of physics, are inserted here for convenience. Both require a very high degree of exhaustion. If a bell is rung in a vacuum, no sound is heard. An electric bell is most convenient for this experiment. It should be carefully supported, so that the sound shall not be transmitted directly to the pump plate. For this purpose it is sometimes hung by threads; a rubber support is also recommended. The experiment is generally more successful, if after exhaustion hydrogen gas is admitted, and the exhaustion repeated. It is, however, almost impossible to destroy all sound. The latent heat of aqueous vapor is well shown by the experiment of freezing water in vacuo. A shallow pan of concentrated sulphuric acid 1 is placed on the pump plate, and on this a wire triangle which supports a flat metallic

1

1 In all cases where sulphuric acid is used to absorb moisture in the presence of metallic surfaces, it should be freed from nitric fumes by boiling it for some time with sulphate of ammonia.

dish holding the water to be frozen. The whole is covered with a small receiver, and exhausted quickly. On removing the pressure from its surface the water is rapidly converted into vapor, which is absorbed by the sulphuric acid as fast as formed; the action therefore continues, the latent heat being obtained at the expense of the water, which accordingly cools until it is converted into ice. By substituting for water more volatile substances, as a mixture of solid carbonic acid and ether, and adding protoxide of nitrogen, the most intense cold yet observed is attained. In the best pumps water may be frozen by its own evaporation, without employing acid to absorb its vapor.

55. MARIOTTE'S LAW.

Apparatus. A modification of Regnault's apparatus may be made chiefly with steam fittings, as shown in Fig. 44. C is a tall mercury gauge formed of glass tubes, connected together by a steam pipe coupling with red sealing-wax. If very high, all the joints must be made like those of Regnault's gauge, but this is unnecessary for pressures below one hundred pounds. A and B are two similar tubes about three feet long, closed above by "petcocks," and attached below by "unions," so that they may be easily removed. E is a reservoir made of 3 inch pipe with caps, to hold the mercury, and with an " valve below, so that it may be emptied if necessary. It is filled by removing the plug in the T at G. I is a small force-pump such as is used in testing gauges, by which water may be drawn from the reservoir K, and forced into E. The water is allowed to flow back by opening the valve H. Remove the plug G, and pour into E enough mercury to fill A, B and C. Work the pump slowly until E is full of water. Then close G and expel any air that may remain by working I and opening H alternately, until no air bubbles rise up through the water in K. Scales are attached to A, B and C, and the first two should be carefully calibrated (Experiment 10). B may be permanently filled with dry carbonic acid, or other gas.

Experiment. A must first be filled with dry air. For this purpose connect it above with a U-tube containing chloride of lime, open the pet-cock and pump up the mercury nearly to the top, thus forcing out the air. Open H a very little, and let the mercury slowly descend. The air is thus drawn into A, first being dried by passing over the lime. Repeat several times to expel all the moisture that may remain, and finally, when full of dry air, close the pet-cock. Read very carefully the height of the mer

cury A, B and C, and record in three columns.

Work the pump

Note

a few times, and take readings at intervals of about 10 inches, until the mercury has nearly reached the top of AC. the height of the barometer and the readings of C and A, when the latter is open to the atmosphere, also the height of the mercury in A when standing at the same level as in C. Write in the 4th and 5th columns the

G

Fig. 44.

D

B

pressure in each case, found by adding to the height of the barometer the difference in level of the mercury columns. In the 6th and 7th columns give the volume of the gas in each case,

deduced from the table of calibration of A and B. Next take the product of the pressure and volume, which would be constant if Mariotte's law were correct, or the volumes inversely as the pressures. Finally, construct curves for the two gases, making abscissas represent these products, and ordinates pressures. These results will be only approximate, owing to the change of temperature the gas undergoes when rarefied or condensed. To diminish this error an interval should be allowed for the gas to attain the temperature of the air of the room, or better, A and B should be surrounded with a water jacket, the temperature carefully noted, and a correction applied.

Much greater accuracy is attained by the following arrangement. A third tube is employed, in the upper part of which two platinum wires pointing upwards are inserted, the volume above them being determined very accurately by inverting, and weighing the mercury required to fill it. This portion of the tube is then enclosed in a larger one, through which water is kept circulating, and its temperature noted by a thermometer. Fill the tube in the usual way with dry gas, then condense it until the mercury is just on a level with the upper platinum point. The mercury in C should now stand near the top of the tube. Open

H, and let the pressure diminish until the mercury in the third tube is exactly on a level with the lower platinum point. Record the pressure in each case. To bring the mercury to the exact level, raise it by the pump and lower by opening H until the point is just perceptible by the slight distortion it produces in the image of objects reflected in the surface of the mercury. Let the pressure diminish to a few inches of mercury, let out a little gas and repeat. The law may be tested for pressures less than one atmosphere by merely drawing off the mercury in C until it stands below that in the other tube. The ratio of the volumes being constant in this experiment, the ratio of the pressures would be its reciprocal, if Mariotte's law were correct. The deviation may be shown by a curve in which abscissas represent the smaller pressure, and ordinates the ratio of the two pressures.

56. GAS-HOLDER.

Apparatus. A good gas-holder containing three, or better, five cubic feet, with scale attached, the bell properly counterpoised, and most important of all, the friction reduced to a minimum. To calibrate it, the standard tenth of a cubic foot of Experiment 19 is employed, and to measure the pressure some very delicate form of gauge should be provided.

D

Experiment. The gas-holder consists of a large bell, AB, Fig. 45, suspended in a circular trough of water, and counterpoised by the weight Fattached to a cord passing over the pulley D. A curved piece of metal E, called the cycloid, is attached to this pulley, and carries a second weight G, which acts at a longer and longer arm as the bell rises. It thus compensates for the diminution of weight of the bell when submerged,

nd renders the pressure nearly the same whether the holder is full or empty. The proper form for E is the involute, a curve in which the perpendicular on the tangent is proportional to the angle described by the radius

A

B

Fig. 45.

vector. The gas is drawn out by a large tube opening into the bottom of the holder at B, and covered at K by a cap with a water seal. A large tube with a stopcock H, also opens out of it, through which the gas may be drawn, or if preferred, through a small tube just below it. To fill the holder with air, remove K and press down on F, when the bell will rise to the top; it may be kept there by replacing K, and closing H. If gas is to be used, it must be filled through II, but this is a much slower process.

A variety of experiments may be performed with this apparatus. First, test the holder. Fill the bell nearly full of air, depress it a little by the hand, let it return, and record the reading of the scale. Then raise it, let it descend, and again read. Repeat several times, and the difference in the results shows the greatest error due to friction. Do the same with other parts of the scale. Next, calibrate the bell. The same method is employed as in Experiment 48, only the air is collected instead of the water. In this case, after emptying the holder, add one tenth of a cubic foot of air at a time, and read the scale after each addition. Repeat drawing out one tenth of a cubic foot from the holder into the glass standard, and see if the readings are the same as before. The great difficulty in this experiment is the change of volume of the air due to changes of temperature. As the bell rises from the water the adhering moisture evaporates, and sometimes lowers its temperature very rapidly. It is, however, customary to assume that the air is saturated with moisture, and at the same temperature as the water with which it is in contact.

Next, measure the pressure for different parts of the scale to see if the compensation is exact. The gauge is attached to a small tube just below H, with an independent outlet from the bell. To save time the pressure may be observed after adding each tenth of a foot in the last experiment. Various forms of gauges may be employed. The simplest is a large U-tube, with a scale attached to each branch. The pressure may thus be determined within a hundredth of an inch. For greater accuracy, a bell glass standing in water may be connected with the holder, and the difference in level of the water within and without it gives the pressure. By using two hook gauges for this purpose great accuracy may be attained. A method in common use is in principle similar to the