He found mental disturbances of some kind in about forty per cent of the cases examined. As a general rule they appeared after the patient had shown evidence of several attacks of the disease. The psychoses occurred in men between the ages of twenty-one and forty and in women between forty-one and sixty. About ninety-five per cent of the mental disorders were to be attributed directly to the effect of the toxin. The remaining five per cent represented individuals with a defective nervous organization or were purely incidental complications. Singer found peculiarities in make-up associated frequently with a predisposition to pellagra. He also expressed the opinion that chronic forms of "insanity" are very rarely caused by the diseases. Sandy[260] made a study of 160 cases at the state hospital at Columbia, South Carolina, in 1916 based on a classification of psychoses quite similar to the one now in use. He found that thirty-five per cent of these belonged to the infective exhaustive group. As a matter of fact, this is the conclusion almost anyone would reach from reading the observations of the earlier writers. These cases were usually characterized by "more or less marked delirium, being accompanied by some confusion and disorientation, there frequently being also hallucinations accompanied by more or less agitation and restlessness." Physically he found, besides well marked symptoms of pellagra, evidences of severe exhaustion, loss of weight, emaciation, fever, sordes, anorexia, and typhoid facies. "In the milder forms of these 'delirious pictures,' as Singer calls them in his contribution to the report of the Thompson-McFadden Pellagra Commission, and as he pointed out, the periods of clouding (of consciousness) may be quite brief and episodic. In such cases in the intervals when the consciousness is practically clear, the general attitude is one of symptomatic depression." Sandy found characteristic manic-depressive forms in eleven per cent of the series reviewed. The depressed types were more common. Here he found retardation of speech and action with a dearth of ideas. In these cases he looks upon pellagra as being merely an exciting etiological factor. The prognosis was not so favorable, however, as it usually is in manic-depressive psychoses, death often being due to the development of central neuritis. In three per cent of the total he found what could only be described as symptomatic depressions, the emotional condition not being so marked as one would expect in the manic-depressive group. In twelve per cent a diagnosis of dementia praecox was made. In these the pellagra was merely an incident and not an etiological factor.