And we now come to other cases, where the percipience has been collective, although it has not been repeated. There is a case[112] where two persons at one moment--a moment of no stress or excitement whatever--see the phantasm of a third; that third person being perhaps occupied with some supraliminal or subliminal thought of the scene in the midst of which she is phantasmally discerned. Both the percipients supposed at the moment that it was their actual sister whom they saw; and one can hardly fancy that a mere act of tranquil recognition of the figure by one percipient would communicate to the other percipient a telepathic shock such as would make _her_ see the same figure as well. The question of the true import of collectivity of percipience renews in another form that problem of _invasion_ to which our evidence so often brings us back. When two or three persons see what seems to be the same phantom in the same place and at the same time, does that mean that that special part of space is somehow modified? or does it mean that a mental impression, conveyed by the distant agent--the phantom-begetter--to one of the percipients is reflected telepathically from that percipient's mind to the minds of the other--as it were secondary--percipients? The reader already knows that I prefer the former of these views. And I observe--as telling against that other view, of psychical contagion--that in certain collective cases we discern no probable link between any one of the percipient minds and the distant agent. In some of that group of collective cases which we are at this moment considering, this absence of link is noticeable in a special way. There is nothing to show that any thought or emotion was passing from agent to percipients at the moment of the apparition. On the contrary, the indication is that there is no necessary connection whatever between the agent's condition of mind at the moment and the fact that such and such persons observed his phantasm. The projection of the phantasm, if I may so term it, seems a matter wholly automatic on the agent's part, as automatic and meaningless as a dream. Assuming, then, that this is so--that these _bilocations_ or self-projections to a point apparently remote from one's body do occur without any appreciable stimulus from without, and in moments of apparent calm and indifference--in what way will this fact tend to modify previous conceptions?