found that lie could do it voluntarily as well as automatically (Fig. 8).
It was easy enough to prove that these communications had nothing to do with spirits. B satisfied himself upon that point in a very short time. But we kept on experimenting, to 
Figs. 4 and 5. determine whether they were of subconscious origin or not. To B himself they felt strangely external. To quote his own words:
"When I wish to write automatically I take a pencil and place my hand upon a sheet of paper. After the lapse of a few minutes I feel a tingling sensation in my arm and fingers; this is followed by a stiffening of the arm and by convulsive movements. After scrawling for a while, it will make a mark which suggests to me the beginning of a letter, and usually the letter will be clearly written almost before the thought enters my mind. It is then followed by some word beginning with that letter, and that by other words, constituting a 'communication' from some 'spirit.' The writing then proceeds quite rapidly. It seems to me that I read it as it is written; sometimes I apparently anticipate the writing, but quite often it does not proceed in accordance with my anticipation. Sometimes the writer seems to be at a loss how to complete his sentence, and begins again. At other times an illegible combination of signs will be repeatedly written, until
finally a word is evolved, and this appears to be what the writer had in mind at the outset. I am now satisfied, however, that there is never any foresight; my hand simply develops the illegible scrawl into the word which I think it most resembles, thus ful-
