Personal Identification and Description 165
typical forms). Unfortunately loops occur in two of Galton's classes only, i.e. as inward and outward loops, and thus the system fails to break up the large class of plain loops which is one of the difficulties of indexing. Further, to be effective, it involves pencilling in the core. However, I feel confident that the choice of a numeral place for each digit and ten classes for each pattern, which need not necessarily be the same for each digit, since the relative frequency of each pattern varies from digit to digit, ought to be the basis of any sound system of finger-print indexing. Here, however, it is the difficulty of breaking up the nearly 50 °/o of plain loops which requires ingenuity and study, and for this Galton could only suggest ridge counting. Breaking up the 25 °/C of "whorls" is a relatively easy matter. Even if we proceed to consider the "nuclei" of the cores (Galton, p. 8) we have five -belonging to the whorls, and only two
9 , & G a ~ I
C7 6 e d e f ,y Fig. 24. "Nuclei" of cores. a, b, c, d, e are "nuclei" of whorls, f and g of loops.
provided by the loops. Thus, considering "inward" and "outward" loops, we have at most four loop classes, and we require greater subdivision. This problem was left unsolved by Galton in this his first memoir (1890), unless his suggestion of ridge-countingbe considered adequate. Our author next proceeds to deal with the identification of patterns. He draws attention to the fact that the patterns may become distorted, either by age or decay, if the times of taking prints are at long intervals. "They may change their shape just as the pattern on different portions of the same piece of machine-made lace may become variously stretched by wear, or shrunk by wet, or even torn" (p. 9). Exactly as we might proceed to identify the lace by counting the threads of corresponding parts of the lace pattern, so we may count the ridges on corresponding parts of two finger-prints. For this purpose Galton uses the axes of reference which I have already discussed.
Besides the counting of ridges Galton also uses the minutiae of which he gives examples in his Fig. 11, but he is careful to warn the reader that two
9 Tz t
Fig. 25. Ambiguous minutiae.
prints of the same finger may show one a fork of ridges and the other a continuous ridge and the terminal of a new ridge, .i.e. one may show a and the other b or c of the above figure. The reason for this is that the ridges