Personal Identification and Description 207 



terminated ridge. It may be applied not only to loops, but to whorls and 

 even to arches to signify that they have an inner curl or hook. 



It must at once be admitted that Galton requires a most imposing battery 

 of additional suffixes and symbols to obtain his Secondary Classification, and 

 further that when this has been accepted and we are able to classify some 

 3000 sets, so that only the slightest difficulty arises in entering or leaving 

 an index, there still remains the fact that difficulties will steadily increase 

 as we mount up from 3000 to 100,000 entries. There may, as Galton himself 

 thought, still be need for three or four anthropometric characters — not for the 

 purpose of identification but for classification. Again, this heavy array of 

 symbols involves much additional work at first in indexing new sets of prints 

 and in reading sets for identification J. When I first read Galton's Secondary 

 Classification and became acquainted with his battery of suffixes, it seemed 

 too unwieldy to be practically applicable, but' a little examination of im- 

 pressions under a lens convinced me that it was reasonably easy for a moderate 

 expert to get a grip upon it. Such experts would be in every "Identification 

 Bureau," and for the mere trained impressor of fingers, such as the prison 

 warder, it is rarely that any necessity arises for reading the prints them- 

 selves. 



More serious defects of Galton's classification are its cumbrous character, 

 and the fact that the letters he uses as subscripts do not convey any hint of 

 their significations. I doubt whether the latter difficulty can possibly be met ; 

 characteristic symbols cannot be found for 20 to 30 subclasses, and if we once 

 realise this, then it does not much matter whether we use numbers or letters 

 provided we use a single one only. If we exclude numbers of ridges in loops, 

 which might be placed as Roman figures in brackets after the index number 

 itself, I believe that ten symbols with powers ought to describe any set of 

 prints. The particular finger — supposing these taken in natural order from 

 left to right — is indicated by the corresponding symbol taken in order from 

 left to right. Now what are the symbols to be? They may be either letters 

 or numbers. At first one might prefer the latter, because if we choose three 

 forms of alphabet, say Greek, Roman and Italic letters, although we can go 

 beyond ten corresponding letters of each, the printed mixture looks clumsy 

 and can only be read out letter by letter. On the other hand, if we use 

 numerals of three types: — say, Roman, Italic and Block — the printed number, 

 while still looking clumsy, if less so, is capable of being read aloud as so many 

 millions, so many thousands, hundreds, tens, etc. The grave disadvantage of 

 the numerical scheme is that it is far less readily adaptable to a written index, 

 where it is not easy to distinguish between Roman, Italic and Block numerals. 

 We shall probably do better therefore to adopt three alphabets, say, the 

 Greek, Italic small and Italic capital. Let us see how this will work. 

 We will first get Galton's symbols d, x, * and f into slightly simpler form. 

 A simple note of interrogation (?) denotes that the print is missing, cannot 

 be taken or is unreadable. A short rule over a letter denotes the print is 



| Of course in about three-quarters of the inquiries it would not be needful to examine the 

 secondary classification at all. 



