At the moment when the table was whirling with a powerful and irresistible rotation, at a given signal we all lifted our fingers. Then keeping our hands united by means of the little fingers, and continuing to form the chain at a height of say an eighth or a quarter of an inch above the table, we continued our circular movement. _To our great surprise the table did the same_; it made in this way three or four turns! We could scarcely believe our good fortune; the by-standers (witnesses) could not keep from clapping their hands. And the way in which the rotation took place was as remarkable as the rotation itself. Once or twice the table stopped following us because the little accidents and interruptions of our march had withdrawn our fingers from their regular distance from the top of the table. Once or twice the table had come to life again--if I may so express myself--when the turning chain had again got into the right relation with it. We all had the feeling that each hand had carried along in its course that portion of the table immediately beneath it. _Seance of September 29_ We were naturally impatient to submit rotation without contact to a new test. In the confusion of the first success we forgot to renew and vary this decisive experiment. When we got to thinking about it afterwards we saw that it behooved us to do the thing over again with more care and in the presence of new witnesses; that it was, above all, important to produce the movement and not merely to continue it, and to produce it in the form of levitations instead of limiting it to rotations. Such was the program of our meeting of September 29. Never was program carried out with greater precision. As a preliminary, we repeated our successful feat of the 24th. While the table was rotating rapidly, the interlocked hands were lifted from it, though continuing to turn above it and form the chain. The table followed, making now one or two revolutions, and now a half or a quarter turn only. The success, more or less prolonged, was certain. We confirmed it several times. But some one might say that, the table being already in motion, the momentum carried it along mechanically while we imagined it was yielding to our fluidic force. The objection was absurd, and we would have challenged anybody to obtain a single quarter of a turn without forming the chain, however rapid might have been the rotation imparted. Above all, would we have challenged anyone to renew its motion when it had been for an instant suspended. However, it is well in such cases to forestall even absurd objections, however little of plausibility they may have. And this particular objection might seem plausible to the inattentive man. It was imperative, then, that we should produce rotation starting from a state of complete inertia.