Automatic writing in its simplest form is merely a variety of the tricks of unconscious action to which, in excited moments, we are all of us prone. The surplus nervous energy escapes along some habitual channel—movements of the hand, for instance, are continued or initiated; and among such hand-movements—drumming of tunes, piano-playing, drawing, and the like—_writing_ naturally holds a prominent place. There is incipient graphic automatism when the nervous student scribbles Greek words on the margin of the paper on which he is striving to produce a copy of iambics. If the paper be suddenly withdrawn he will have no notion what he has written. And more, the words written will sometimes be _imaginary_ words, which have needed some faint unconscious choice in order to preserve a look of real words in their arrangement of letters. A complete graphic automatism is seen in various morbid states. A man attacked by a slight epileptiform seizure while in the act of writing will sometimes continue to write a few sentences unconsciously, which, although probably nonsensical, will often be correct in spelling and grammar. Again, in the case of certain cerebral troubles, the patient will write the _wrong_ word—say, “table” for “chair;”—or at least some meaningless sequence of letters, in which, however, each letter is properly formed. In each of these cases, therefore, there is graphic automatism. And they incidentally show that to write words in a sudden state of unconsciousness, or to write words against one’s will, is not necessarily a proof that any intelligence is at work besides one’s own. Still further; in spontaneous somnambulism, the patient will often write long letters or essays. Sometimes these are incoherent, like a dream; sometimes they are on the level of his waking productions; sometimes they even seem to rise above it. They may contain at any rate ingenious manipulations of data known to his waking brain, as where a baffling mathematical problem is solved during sleep. From the natural or spontaneous cases of graphic automatism let us pass on to the induced or experimental cases. I will give first a singular transitional instance, where there is no voluntary muscular action, but yet a previous exercise of expectant attention is necessary to secure the result.