During his experiments he had found that platinum became exceedingly hard after it had been heated several times to incandescence by current flowing through it. This apparently raised its melting temperature so he was able to increase the operating temperature and therefore greatly increase the candlepower of his lamps after they had been heated a few times. Examination of the platinum under a microscope showed it to be much less porous after heating, so he reasoned that gases were occluded throughout the platinum and were driven out by the heat. This led him to make a lamp with a platinum wire to operate in vacuum, as he thought that more of the occluded gases would come out under such circumstances. [Illustration: EDISON’S CARBON LAMP OF OCTOBER 21, 1879. This experimental lamp, having a high resistance carbon filament operating in a high vacuum maintained by an all-glass globe, was the keystone of Edison’s successful incandescent lighting system. All incandescent lamps made today embody the basic features of this lamp. This replica is in the Smithsonian Institution exhibit of Edison lamps. The original was destroyed.] These lamps were expensive to make, and, knowing that he could get the requisite high resistance at much less cost from a long and slender piece of carbon, he thought he might be able to make the carbon last in the high vacuum he had been able to obtain from the newly invented Geissler and Sprengel mercury air pumps. After several trials he finally was able to carbonize a piece of ordinary sewing thread. This he mounted in a one-piece all glass globe, all joints fused by melting the glass together, which he considered was essential in order to maintain the high vacuum. Platinum wires were fused in the glass to connect the carbonized thread inside the bulb with the circuit outside as platinum has the same coefficient of expansion as glass and hence maintains an airtight joint. He reasoned that there would be occluded gases in the carbonized thread which would immediately burn up if the slightest trace of oxygen were present, so he heated the lamp while it was still on the exhaust pump after a high degree of vacuum had been obtained. This was accomplished by passing a small amount of current through the “filament,” as he called it, gently heating it. Immediately the gases started coming out, and it took eight hours more on the pump before they stopped. The lamp was then sealed and ready for trial. [Illustration: DEMONSTRATION OF EDISON’S INCANDESCENT LIGHTING SYSTEM. Showing view of Menlo Park Laboratory Buildings, 1880.]