This study examined the reliability and validity of graphology as an assessment tool in selecting candidates for an infantry officers' training course. Six professional graphologists' predictions were compared with those of 6 psychologists and 6 laymen who evaluated the same 65 handwriting specimens on 13 scales. Coefficients of concordance in the graphologists' group were relatively low and ranged from .20 to .37 but were higher than those of psychologists (.17 to .36) and laymen (.09 to .20). The median validity coefficients of graphologists in predicting success in the course ranged from .11 to .26, while the validities of psychologists were from .07 to .23 and of laymen from —.05 to .17. The median validities of a General Evaluation Scale were .26, .20, and .11 for graphologists, psychologists, and laymen, respectively. The possible uses of graphology in selection and direction of research were discussed.