Personal Identification and Description   .183

While convinced that the chance of two individuals actually possessing the same finger-print in all its minutiae is infinitesimally small-as small as the chance that two woodcutters given the same topic would produce two blocks identical in every line and dot-yet one recognises that Galton's treatment, however ingenious, lacks the power of compelling conviction. Nature probably works more definitely to form a whole pattern than can be mimicked by Galton's 24 "independent variable" squares. He himself writes that

"it is hateful to blunder in calculations of adverse chances, by overlooking correlations between variables, and to falsely assume them to be independent, with the result that inflated estimates are made which require to be proportionately reduced. Here, however, there seems to be little

room for such an error." (p. 109.)

It is the last sentence only we would call in question. After all it is the minutiae, rather than the pattern, by which identification is determined. Hence we might consider the problem as follows : These minutiae are not points, the ridges having a measurable thickness. Let us suppose a ridge-interval square to cover the area within which, if two such minutiae occurred in two -prints under comparison, we should hold these minutiae to be identical in position. Galton's 6-ridge interval squares contain 36 little 1-ridge interval squares, and the chance of a given minutiae occurring in one of these

is 36, say 25 roughly. Now Galton takes 24 such squares to a finger-print,

and roughly there are 20 to 30 or even more minutiae in a print, say one to each 6-ridge interval square; then the probability that the minutiae will be placed each in its right compartment in its 6-ridge interval square is less

1 24   1

than (25) , i. e. less than 2120 .- Actually it is considerably less than this because

although the minutiae do not tend to cluster each one of them is not confined to its own 6-ridge interval square. Further all minutiae are not alike, e.g. ridge terminals. I think we may suppose a far more random, that is, less correlated, distribution of minutiae, than of parts of a pattern, and still conclude with Galton that it is very unlikely that two persons in the universe have the same print on any digit, as judged by its minutiae, still less on all ten digits.

Galton concludes this chapter characteristically as follows

"We read of the dead body of Jezebel being devoured by the dogs of Jezreel, so that no man might say, 'This is Jezebel,' and that the dogs left only her skull, the palms of her hands, and the soles of her feet; but the palms of the hands and soles of the feet are the very remains by which a corpse might be most surely identified, if impressions of them during life were available." (p. 113.)

Chapter VIII (pp. 114-130) -is entitled Peculiarities of the Digits. The data Galton uses in this chapter are the prints of the ten digits of 500 different persons. His objects are twofold : (i) - to find the association of particular patterns with the individual digits, and (ii) to determine, if a particular digit has a given pattern, what is the chance that any other digit will have the same pattern. In discussion of these problems Galton uses only the triple