Bessler might have easily put together a deception of this kind at his own house, but to do so at the Castle Weissenstein would have involved various architectural problems, and posed a great risk of detection.

By all accounts, those who inspected the machine always went away satisfied that the bearings were open and free from any suspicion. Hundreds of competent, honest, impartial people carefully examined every aspect of the machine, looking specifically for the slightest sign of fraud, and none was ever found.

The most obvious answer is that there was a man in it. Either straddling the axle or sitting in some sort of saddle, he could have easily provided the power to rotate the wheel and perform the work through a peddle gear or directly, as in a treadmill.

This solution, however, poses more problems than it solves. The early wheels were just a few inches wide and between three and five feet in diameter. Supposing there was a man or child in the larger wheels, it is unlikely that he could exist for two months in a sealed room during the Kassel examination and leave no sign of his presence.

It is true that both weight driven and spring driven clocks have been made which will go for long periods without rewinding. But it is also true that the stored power necessarily must be doled out in such infinitesimal doses that all the moving parts must be kept as light and frictionless as possible; hence such clocks are incapable of doing any more work than keeping themselves going.

If we assume that Bessler's largest wheel weighed a few hundred pounds, it is questionable that enough power could have been stored inside of it in the form of weights or springs to keep it spinning at a constant speed for more than fifty days, as it did in the Kassel test.

Even if we grant the possibility of having done so, there are two other problems with the clockwork theory. The power available would have been so slight that the wheel certainly could not have done work. When a load was applied to the axle, it would have slowed down gradually and steadily until it stopped altogether. Instead, applying a load merely reduced the wheel's constant speed by a few revolutions per minute.

It is also hard to imagine how a clockwork mechanism could have delivered enough power to account for the well documented quick acceleration of the wheel. With the limited power available, a wound spring would have taken some time, probably measured in hours, to work up to a speed of 26 revolutions per minute. In fact, after barely being started with two fingers, the wheel reached that speed in just one turn.