Mendillo and Baumgardner first observed the moon when it was only a quarter full--the light of a full moon would obscure the faint atmosphere--and noticed that the atmosphere was most dense in the direction facing the sun. This allowed them to rule out meteors as the source of the atmosphere. Meteors, they argue, would be expected to hit the moon from all directions, thereby creating a fairly uniform atmosphere.
Observations made during a lunar eclipse ruled out the solar wind. An eclipse occurs when the moon passes through Earth’s shadow, or umbra, while its atmosphere is still lit by the sun, and is the best time for photographing the atmosphere. On November 29, 1993, Mendillo and Baumgardner used a mask on their telescope to block residual moonlight visible even during an eclipse. They detected sodium atoms twice as far out as anyone had before--nearly 10,000 miles from the moon.
Moreover, they made those observations when the moon was inside Earth’s magnetic field. The field is swept into a long tail on the planet’s nightside by the charged particles of the solar wind, but it also deflects that wind. Before the eclipse, the moon had been inside the magnetotail and protected from the solar wind for several days, and yet its atmosphere was undiminished. But it was battered by sunlight--which is unaffected by the magnetic field--for all but a few hours during the eclipse. Mendillo and Baumgardner concluded that sunlight must be primarily responsible for creating the lunar atmosphere.
Sunlight is constantly dislodging atoms from the lunar surface, Mendillo explains, because the atoms are only loosely bound to the lunar dust. Some of the atoms drift off into space, never to return, but most of them slow and eventually fall back to the moon. It’s not a permanent atmosphere in the sense ours is, says Mendillo. It is continually being created and lost.


