It is necessary to point out that the existence of the department is not so much for the purpose of detecting crime as of detecting criminals. In the administration of justice a judge takes the past career of a prisoner into consideration when passing sentence. The main work of the department is to furnish the clue to a past career by scrutinising the finger-prints of persons on remand to discover whether they are habitual criminals or not. A thousand aliases will not help a man, no change of appearance, no protestations of mistake, if his prints correspond with those in the files. But it is all so simply done. There is nothing spectacular, nothing imposing about the process. Practically all that is needed is a piece of tin, some printer's ink, and a sheet of paper. Within a few minutes afterwards his record can be known. Compare this with the old Bertillon system of anthropometric measurements. Bertillon's system depends on the fact that after a person reaches maturity certain portions of the body are always the same in measurement. The theory is sound, but the difficulties in the way of applying it are immense. In his book Sir Edward Henry has pointed out the defects of the system. The instruments are costly, measurers have to be specially trained, and even so may make a mistake--an error of two twenty-fifths of an inch will prevent identification--the search among the records may take an hour or more, and, moreover, through carelessness or inattention, the whole data may be wrong. For six years--from 1895 to 1901--this system was in force at Scotland Yard. The maximum number of identifications in any one year was 500. In 1913, by the aid of finger-prints, 10,607 persons were identified. Roughly, it is all a matter of classification into "arches," "loops," "whorls," and "composites." It is intricate to describe, but simple to carry out. To the uninitiated it inevitably suggests the old problem "think of a number, double it--." What happens is this: Every print for primary classification purposes is considered as a loop or a whorl. The fingers are taken in pairs and put down something like this: L. L. W. L. L. -------------------- L. W. W. W. W. Now a whorl occurring in the first pair would count sixteen, in the second, eight, and so on. The loops are ignored. Consequently, the number in the above formula is: 0. 0. 4. 0. 0. ---------------- 0. 8. 4. 2. 1. These are added together and become 4-15. The figure 1 is added above and below, and the searcher knows that he has to look for the record he wants in the sixteenth file of Number 5 horizontal row in a cabinet specially arranged. Of course, sub-classification is carried much farther than this, but it is scarcely necessary to elaborate the point.