Graphology, the theory and practice of diagnosing character from handwriting, is gradually developing from its initial stage, in which it had been the mysterious art of a few people who claimed to be especially gifted, into a science which can be taught and studied. The widespread mistrust of it is partly justified by this state of development. From a scientific point of view it needs more precision of method and terminology as well as corroboration of its results by objective methods. The description of the character as a whole and assessment of the significance of single features still depend to a large extent on the graphologist's experience and knowledge of human nature as well as on his stylistic abilities, and some other subjective factors. Moreover, only part of the accepted graphological interpretations have been based on inductive methods using statistical comparison, experiment and the like; for the rest, they have regarded handwriting as the more or less unwitting expression of the subject's personality, and they have attempted to assess this either by empathy, or by an effort to understand and explain which might itself be founded on the graphologist's private philosophy or on the psychoanalytical system. The results obtained in this way called for corroboration by statistical evidence. The present writer therefore welcomed Dr. Eysenck's suggestion to submit her graphological reports to objective control in the following way: She was to answer a questionnaire on the temperamental qualities of 50 neurotic patients, a sample of whose handwriting was given her, assess their intelligence, give a short character description, and, last of all, match the specimens of handwriting against anonymous summaries of the case-records of the patients. Her results were then compared with the patients' own answers to the questionnaire, their score in the Progressive Matrices Test, and the psychiatrist's description of the patient's character. Dr. Eysenck has reported the numerical results. In the present paper the validity of the graphological opinion on personality traits confirmed by at least one of the sources of control will be examined, and the handwriting signs of these traits will be set forth.