Paddlewheel

Crookes' paddlewheel tube, from his 1879 paper On Radiant Matter
Crookes put a tiny vaned turbine or paddlewheel in the path of the cathode rays, and found that it rotated when the rays hit it. The paddlewheel turned in a direction away from the cathode side of the tube, suggesting that the rays were coming from the cathode. Crookes concluded at the time that this showed that cathode rays had momentum, so the rays were likely matter particles. But in 1903, J. J. Thompson proved that the paddlewheel wasn't turned by the force of the cathode rays hitting it, but by the radiometric effect. When the rays hit a paddle, they heated the side they hit. The air next to that side of the paddle expanded, pushing the paddle away. All this experiment really showed was that cathode rays could heat objects.
http://en.wikipedia.org/wiki/Crookes_tube