the disagreeable compulsion of the question than the invention of an answer. One should therefore interrogate no more than is absolutely necessary and should formulate his questions as "unsuggestively" as possible.
c. Since memory is such an important function it is natural to ask whether we must rest satisfied with its demonstrated imperfection. Is it not possible to secure an improvement by pedagogical means? The question is to be answered affirmatively; just as observation (Anschauung) can be systematically cultivated, so can memory. The improvement is subject to experimental demonstration; experiments repeated on the same children (each time of course with a new picture) showed a clear improvement. The chief educative effect was in this case due to self-correction. After the making of the report the picture was shown again to the child and he was required himself to discover the errors which he had made. Such exercises of memory may be scattered through the work upon any school topic as opportunity offers.
7. Practical Consequences for Law. (The consequences here mentioned have reference in the first instance to German jurisprudence and court procedure. To what extent analogous points of view hold also for American conditions those familiar with the latter must decide. )
a. The first and obvious consequence of the psychology of testimony is a negative one, a diminution of the reliance which is to be placed in the reports of witnesses. The notion, still tolerably prevalent, that the faithfully sworn testimony of a mentally competent witness is in general to be regarded as an exact presentation of reality, is without justification. In Germany the new view has already caused the testimony of children especially to be less highly valued than formerly. It would be a mistake, however, to ascribe to the psychology of testimony destructive consequences only; its positive consequences are still more important.
b. The examining officer is able by the manner of his questioning to predetermine in a measure the degree of the erroneousness of the testimony. The more he leaves to spontaneous narration, and the less "suggestive" his questions, the less will be the danger of falsification.
c. When identification is necessary the witness should make it, whenever possible, by choice from a group of similar persons or things (Wahlkonfrontation) and not by indicating whether a single individual presented to him is the one in question (Einzelkonfrontation), because of the powerful "suggestive' ' effect of the latter procedure.
d. Psychological experiment shows what degree of confidence ought in general to be placed in particular classes of
