Early in working with our cosmopolitan population, (and to a less extent the trouble would have shown itself among those of one language) we saw that any method of mental evaluation which was based largely upon language tests, whether or not given by such a questionaire method as that of Binet was quite unfair. Language, our universal method of communication, does not cover all the social graces, all the social values, nor does adeptness in the use of it mean unusual general ability. In fact, nothing has been any more striking in our findings than that some otherwise normal individuals have special defects for language, and that some general defectives have such powers in manipulating words that pass everywhere, even in courts, as normal and even bright. It was soon felt that over and beyond tests which involve the medium of communication, it would be more profitable to observe a performance with concrete material which possibly might be arranged to measure some of the socially most valuable mental qualities. Such performance tests have had a great growth in these five years, as emanating from several centers, until at present they are quite widely used. For work with offenders, there is at present a wide range of tests to choose from. For practical clinical work even more tests are desirable, and it may be that some of those now used will be gradually discarded and replaced by others. It stands out very clearly that what one would like to know particularly about offenders is how they grade, not only in general intelligence, but also on tests which may possibly evaluate the powers of apperception, of mental representation, of self-control, of the ability to learn by experience, and so on. Defects along these lines seem likely to stand in much greater correlation to delinquent tendencies than any other we could name. Special educational and industrial disabilities may make for social failure and so may indirectly lead also to delinquency. It is usually not difficult to ascertain the nature of these defects. A few steps towards discovery of vocational aptitudes can also be made by the use of tests. We still think that the early advice to keep our testing methods simple and direct was thoroughly sound. It is evident that we can use with scientific safety the Binet scale for grouping young children and defectives up to the level of 10 years. Beyond this, by using a wide range of other tests, we can discriminate other subnormal groups who are either defective in general or in special abilities.