The greater number of the tests had to do with more strictly mental characteristics. Poincare had no visual images or memories, except in the transition state between waking and sleeping, when he had frequent visual hallucinations of remarkable distinctness. In his waking life he relied chiefly on motor images and tendencies, thinking of geometrical forms in terms of optical or manual movements. He had no visual "schemes," but represented time, in his thinking, by a rotation of the eyes on their axes. In his youth he had pronounced colored hearing, which was evoked not by the form but by the sound of letters and words. He had no other synesthesias. Tests of recognition memory for length of lines, reproduction of drawings seen once, etc., are said to have shown exceptional memory capacity. The memories were held with the aid of motor imagery, and the reproduction was often not from the image but on the basis of an analysis of the material which had been presented to him. He had a memory span for digits of about eleven, as compared with the ordinary record of about eight. In the case of letters he had an auditory memory span of nine, and a visual span of seven. Mechanical memory did not seem to be particularly good, and much emphasis is laid on Poincare's tendency to use memory devices when remembering this non-logical material; he employed analysis and incidental schemes whenever possible. He had a "remarkable facility in mental calculation," which is said not to be the rule with mathematicians. In tests of logical memory he was superior to both Zola and Dalou, and here again his memory was found to be analytical and artificial rather than mechanical. All material was arranged in a coherent scheme or system, and it was this system, rather than the material, that was remembered.