In 1907 Dreyfus,[284] at that time an assistant of Kraepelin's, made an elaborate study of the cases previously diagnosed as involution melancholia in the Heidelberg clinic. During a period of fourteen years, a total of seventy-nine were reported. A thorough investigation by Dreyfus showed that two-thirds of these had made complete recoveries or improved to such an extent as to be able to go home. Only eight per cent showed a marked mental deterioration. He also found that over half of the series had more than one attack, usually depressions. One-third of the patients died and were thus eliminated from further consideration. The duration of the attack was over three years in one-third of the cases reviewed. Fifteen per cent recovered in from three to five years, nine per cent in from six to eight years, and eight per cent in from ten to fourteen years. He was of the opinion that after a careful study of the hospital records the symptoms found could all be explained on the basis of manic-depressive insanity, usually of a mixed form. Kraepelin had reported that forty-nine per cent of his cases deteriorated mentally. Dreyfus reduced this on further observation to only eight per cent. On analysis he found, in many instances, brief periods of manic elation, sometimes only a matter of hours or a few days, evidences of excitability, manic suggestion in the eagerness of the patient to communicate his troubles to others, and inhibitory processes indicated by a lack of interest, loss of affection or even difficulty of thinking. Dreyfus concluded that the depressions of late years were not so common as had been supposed and that a sufficient knowledge of their history showed that they had usually exhibited previous attacks. He thought that the long duration of the disease probably led to erroneous ideas as to its termination in deterioration.