Since there's no air to move the loudspeaker didn't "make any sound" yet was still oscillating at 1 kHz. Reintroducing air into the bell jar allowed for air to be moved, thus "creating sound."
I suspect that it's a shielding problem with the microphone picking up stray fields from the audio amplifier. It's really a pretty poor experiment. Not explained what they were doing or why the results. The o'scope is worthless and there's not any correlation between the gauge and the sound intensity. Sadly, fairly typical of a school lab demonstration from students who are just memorizing their lessons. I agree witht he commenter that thought the microphone should have been inside the bell jar.
