afflictions of a large number of the victims of "shell-shock." But, whatever be the cause, the patient is reduced mentally to an elementary state, resembling that of a child, and needs re-educating in the elements of language. Now, from his careful study of the pathological phenomena, manifested in these cases, Bergson draws some very important conclusions in regard to the nature of Memory and its relation to the brain. In 1896, when he brought out his work Matiere et Memoire, in Paris, the general view was against his conclusions and his opinions were ridiculed. By 1910, a marked change had come about and he was able to refer to this in the new introduction.[Footnote: See Bibliography, p. 158.] His view was no longer considered paradoxical. The conception of aphasia, once classical, universally admitted, believed to be unshakeable, had been considerably shaken in that period of fourteen years. Localization, and reference to centres would not, it was found, explain things sufficiently.[Footnote: The work of Pierre Janet was largely influential also in bringing about this change of view.] This involved a too rigid and mechanical conception of the brain as a mere "box," and Bergson attacks it very forcibly under the name of "the box theory." "All the arguments," he says, "from fact which may be invoked in favour of a probable accumulation of memories in the cortical substance, are drawn from local disorders of memory. But if recollections were really deposited in the brain, to definite gaps in memory characteristic lesions of the brain would correspond. Now in those forms of amnesia in which a whole period of our past existence, for example, is abruptly and entirely obliterated from memory, we do not observe any precise cerebral lesion; and on the contrary, in those disorders of memory where cerebral localization is distinct and certain, that is to say, in the different types of aphasia, and in the diseases of visual or auditory recognition, we do not find that certain definite recollections are, as it were, torn from their seat, but that it is the whole faculty of remembering that is more or less diminished in vitality, as if the subject had more or less difficulty in bringing his recollections into contact with the present situation." [Footnote: Matter and Memory, p. 315 (Fr. pp. 264-265).] But as it is a fact that the past survives under two distinct forms, viz., "motor mechanisms" and "independent recollections," we find that this explains why "in all cases where a lesion of the brain attacks a certain category of recollections, the affected recollections do not resemble each other by all belonging to the same period, or by any logical relationship to one another, but simply in that they are all auditive or all visual or all motor. That which is damaged appears to be the various sensorial or motor areas, or more often still, those appendages which permit of their being set going from within the cortex rather than the