We must, of course, pay for electricity according to the cost of producing it. To produce .5 ampere at 110-volt pressure costs the same as one ampere at 55-volt pressure, or .25 amperes at 220 volts. It will be noticed that the products of the two factors in each case are the same. The product of an ampere multiplied by a volt is a watt. In each of the above three cases the amount of electrical energy is 55 watts. This will produce a definite quantity of light--about 16 candle-power when the carbon filament is used, and this quantity does not vary as either volts or amperes, but as the product of these, namely, watts. Each of these lamps is called a 55-watt lamp, and, since they each give 16 candle-power of light, a carbon filament lamp gives one candle-power of light for three and a half watts of electricity. Electricity for lighting purposes usually costs _10 cents per kilowatt hour_, that is, 10 cents for 1000 watts for one hour, or one cent for 100 watts for one hour. Hence a 55-watt lamp costs a trifle more than half a cent for one hour, or exactly .55 cents, and a 32-candle-power lamp costs 1.1 cents per hour. We introduced into the socket a 48-candle-power 110-volt tungsten lamp (Fig. 97), and turned on the 110-volt current. The ammeter showed 55 ampere. Hence the lamp is a 60-watt lamp, and requires one and a quarter watts per candle-power. That is, the metal tungsten is nearly three times as efficient as carbon for producing light from electricity. [Illustration: Fig. 97] With pincers we broke off the tip of a 32-candle-power carbon filament lamp, making a small hole in the large end of the bulb. The air rushed in. We then put the lamp in the socket and turned on the current. The carbon filament glowed as usual, and slowly burned up, growing smaller as it did so. The ammeter which was in circuit showed that the current, which was one ampere at the beginning, grew steadily less as the filament grew smaller, until finally when it was about one quarter of an ampere, the circuit was broken by the filament burning in two. We removed the lamp from the socket and with a dropper tube introduced a little lime water, and shook it to absorb any gas which might have been formed in there. It became milky white, as it always does when introduced where carbon has been burned. This would be a sufficient proof that the filament was made of carbon, if we did not already know it. The air is exhausted from these bulbs to prevent the carbon filament from burning up. [Illustration: Fig. 98]