The third form of procedure is full of all sorts of difficulties and sources of error, many of which are, at the present stage of our knowledge, irremediable. In selecting a new test which will involve the same mental attitude and call for the exercise of the same psychological functions as are needed in the work itself, we are handicapped by the unreliability of the introspection of the examinee and also by our inadequate ability to recognize, identify and classify psychological functions even when we are confident that these are present. The statement of motormen that the manipulation of a crank in connection with a strip of checkered paper makes them feel quite as they do when guiding their cars through a crowded thoroughfare is far from a guaranty "that the mental function which they were going through had the greatest possible similarity with their experience on the front platform of the electric car." It is much more conceivable that the "mental attitude" referred to was merely the vague feeling that "Something is happening now," "This keeps me busy," or "What a nuisance this thing is." And even if we knew the mental functions involved, as would be demanded by the vocational psychograph method, we are still a long way from the time when we can exhibit even a single psychological test and state just what function or functions its performance does or does not, may or may not, involve. Indeed we do not even know what the various distinct mental functions are, or whether, as a matter of fact, there are such distinct functions. After all, the miscellaneous, random, and purely empirical method of Lough, Lahy and Woolley seems to be the most promising experimental procedure for the immediate present, and perhaps for some time to come. This method is, to be sure, but a rough, provisional and unanalyzed expedient. It calls for long and patient cooperative labor. It does not at once afford us the systematic scientific insight which we may wish we possessed. But it will at least save us from the delusion that we already possess such insight, and it should serve to check the fervent and semi-religious zeal that leads us to mistake prophecy for service. Analysis and classification of the results which this method yields are possible when the results are accumulated in adequate measure.