It could have been done by drawing the chair with a string or by the medium putting her foot sufficiently far out. But it happened over and over again (five or six times), automatically moving, and that so violently that the chair jumped about the floor in a topsy-turvy fashion and ended by falling bottom side up without anybody having touched it.
4. Here is a fourth case re-observed this year, after having been several times verified by me, notably in 1898.
Curtains near the medium, but which it is impossible for her to touch, either with the hand or the foot, swell out their whole length, as if inflated by a gusty wind. I have several times seen them envelop the heads of the spectators as if with cowls of Capuchin monks.
5. Here is a fifth instance, authenticated by me several times, and always with the same care.
While I am holding one hand of Eusapia in mine, and one of my astronomical friends, tutor at the Ecole Polytechnique, is holding the other, we are touched, first one and then the other, upon the side and on the shoulders, as if by an invisible hand.
The medium usually tries to get together her two hands, held separately by each of us, and by a skilful substitution to make us believe we hold both when she has succeeded in disengaging one. This fraud being well known by us, we act the part of forewarned spectators, and are positive that we have each succeeded in holding her hands apart. The touchings in this experiment seem to proceed from an invisible entity and are rather disagreeable. Those which take place in the immediate vicinity of the medium _could_ be due to fraud; but to some of them this explanation is inapplicable.
This is the place to remark that, unfortunately, the extraordinary character of the phenomena is in direct ratio with the absence of light, and we are continually asked by the medium to turn down the gas, almost to the vanishing point: "_Meno luce! meno luce!_" ("Less light, less light"). This, of course, is advantageous to all kinds of fraud. But it is a condition no more obligatory than the others. There is in it no implication of a threat.
We can get a large number of mediumistic phenomena with a light strong enough for us to distinguish things with certainty. Still, it is a fact that light is unfavorable to the production of phenomena.