There are also attacks which ostensibly consist only of motor phenomena and lack the passionnelle phase. If it is possible during such an attack of general twitching, cataleptic rigidity or an _attaque de sommeil_, to put one’s self _en rapport_ with the patient, or still better, if one succeeds in evoking the attack in a hypnotic state, it will then be found that here, too, the root of it is the memory of a psychic trauma, or of a series of traumas which make themselves otherwise prominent in an hallucinatory phase. A little girl had suffered for years from attacks of general convulsions which could be and were taken for epileptic. She was hypnotized for purposes of differential diagnosis and she immediately merged into one of her attacks. On being asked what she saw she said, “The dog, the dog is coming,” and it was really found that the first attack of this kind appeared after she was pursued by a mad dog. The success of the therapy then verified our diagnosis. An official who became hysterical as a result of ill treatment on the part of his employer suffered from attacks, during which he fell to the floor raging furiously without uttering a word or displaying any hallucinations. The attack was provoked in a state of hypnosis and he then stated that he lived through the scene during which his employer insulted him in the street and struck him with a cane. A few days later he came to me complaining that he had the same attack, but this time it was shown in the hypnosis that he went through the scene which was really connected with the onset of his disease; it was the scene in the court room when he was unable to get satisfaction for the ill treatment which he received, etc. The memories which appear in hysterical attacks or which can be awakened in them correspond in all other respects to the causes which we have found as the basis of the continuous hysterical symptoms. Like these they refer to psychic traumas which were prevented from alleviation by ab-reaction or by associative elaboration, like these they lack entirely or in their essential components the memory possibilities of normal consciousness and appear to belong to the ideation of hypnoid states of consciousness with limited associations. Finally they are also amenable to therapeutic proof. Our observations have often taught us that a memory which has always evoked attacks becomes incapacitated when in a hypnotic state it is brought to reaction and associative correction. The motor phenomena of the hysterical attack can partly be interpreted as the memory of a general form of reaction of the accompanying affect, or partly as a direct motor expression of this memory (like the fidgeting of the whole body which even infants make use of), and partly, like the hysterical stigmata—the continuous symptoms—they are inexplainable on this assumption.