[Illustration: _Magnetic tapes of special codes using notched ones in place of zeros being prepared for use._] But the real test was yet to come. Special modified signals for two of the fifteen Telstar commands, using our new notched ones in place of the usual zeros, were put on magnetic tape (_see photograph_). Then, on December 20th, when Telstar made its 1492nd pass over Andover, Maine, a group of tired engineers huddled about the mass of equipment they had assembled. Finally, on the third try, the notched pulses were successful; Telstar’s telemetry flashed back the word that the proper relay had operated upon command. Removing the Ionization We now wanted to get Telstar to do something that had seemed to work in the laboratory. The transistors most affected by radiation were those operating under continuous reverse bias, to whose surfaces unwanted ions were attracted. If we removed the voltage from these transistors, we felt that the ionization layer would be dissipated, and they would act normally again. Our plan was to prepare a complete taped program of all fifteen commands, and carefully disconnect Telstar’s storage battery (using command SS). Then, when the satellite went into eclipse, there would be no power available from the solar cells either, and—if our calculation was right—the complete lack of voltage ought to restore the transistors to working order. This was a hazardous procedure, for if something went wrong we might have a completely silent satellite on our hands. As it turned out, an accident did happen—but one of a different and much more fortunate kind. On December 27th Telstar misinterpreted our “trick” commands and disconnected its own battery before we asked it to. Then, as the satellite went into the earth’s shadow, we held our breath while all its power stopped and the telemetry went silent. But, as we had hoped, a rest period with all power removed from the deteriorated transistors apparently made them work almost normally once again. On January 1, 1963, we were able to disconnect the battery in regular fashion—that is, using the one-and-zero code. After this was done, and all power had been removed, both decoders again would operate when given normal commands (actually, the first one restored to duty was decoder No. 2, which had gone out of order first, back in August). Back to Normal—For a Time For more than a month Telstar I behaved as it should, and our communications experiments, including television broadcasts, were resumed on January 3rd. During this time we used both normal commands and our special notched-pulse modified commands. Whenever normal commands became intermittent we used the modified commands to disconnect the battery for several eclipses.