Pondering over this, Hauksbee tried various experiments, revolving pieces of amber, flint, steel, and other substances in his exhausted air-pump receiver, with negative, or unsatisfactory, results.

Finally, it occurred to him to revolve an exhausted glass tube itself.

Mounting such a globe of glass on an axis so that it could be revolved rapidly by a belt running on a large wheel, he found that by holding his fingers against the whirling globe a purplish glow appeared, giving sufficient light so that coarse print could be read, and the walls of a dark room sensibly lightened several feet away.

As air was admitted to the globe the light gradually diminished, and it seemed to him that this diminished glow was very similar in appearance to the pale light seen in the mercurial barometer.
Could it be that it was the glass, and not the mercury, that caused it?
Going to a barometer he proceeded to rub the glass above the column of mercury over the vacuum, without disturbing the mercury, when, to his astonishment, the same faint light, to all appearances identical with the glow seen in the whirling globe, was produced.
Turning these demonstrations over in his mind, he recalled the well-known fact that rubbed glass attracted bits of paper, leaf-brass, and other light substances, and that this phenomenon was supposed to be electrical.

This led him finally to determine the hitherto unsuspected fact, that the glow in the barometer was electrical as was also the glow seen in his whirling globe.