June ii, 1891] 



NATURE 



41 



Elasmobranchs, more particularly to the Chimasroids on the one 

 hand, and to such an ancient Selachian type as Clilamydoselache 

 on the other ; but, at the same time, the Ganoids probably arose 

 from the common ancestral stock not very far off. Though 

 retaining many primitive characters, the Dipnoi, and more 

 especially Protoptertis and Lepidosiren, are in some respects 

 hijjhly specialized, the specialization being largely due to a 

 change of habit. 



"Method of Indexing Finger Marks." By Francis Gallon, 

 F.R.S. 



Sufticient proof was adduced by me in a memoir read Nov- 

 ember 27, 1890, before the Royal Society (Phil. Trans., B, 

 1891), of the extraordinary persistence of the papillary ridges 

 on the inner surface of the hands throughout life. It was shown 

 that the impression in ink upon paper of each finger tip, con- 

 tained on the average from twenty-five to thirty distinct points 

 of reference, every one of which, with the rarest exception, 

 appeared to be absolutely persistent. Consequently that it was 

 possible to affirm with practical certainty whether or no any two 

 submitted impressions were made by the fingers of the same 

 person. 



In the present memoir I shall explain the way in which finger 

 prints may be indexed and referred to after the fashion of a 

 dictionary, and on the same general principle as that devised by 

 A. Bertillon with respect to anthropometric measures, whose 

 ingenious method is now in regular use on a very large scale in 

 the criminal administration of France and elsewhere. I desire 

 to show how vastly the practical efficiency of any such method 

 as that of A, Bertillon admits of being increased by taking 

 finger prints into account in the way to be described. 



It must not, however, be supposed that the use of indexing 

 finger marks is limited to the above purpose, the power of doing 

 so being equally needed for racial and hereditary inquiries. I 

 do not dwell upon these applic;ations now, simply because I am 

 engaged in making them, and the results are not yet ready to be 

 published. I ought, however, to mention that a great increase 

 of experience has fully confirmed my earlier views, that finger 

 marks are singularly appropriate subjects of anthropometric 

 study owing to many distinct reasons. The impressions are 

 easily to be made by anyone who has the proper appliances at 

 hand. They are as durable as any other printed matter, and 

 they occupy very little space. The patterns are usually sharp 

 and clear, and their miniiticB are independent of age and growth. 

 They are necessarily trustworthy, and no reluctance is shown in 

 permitting them to be taken, which can be founded either upon 

 personal vanity or upon an unwillingness to communicate un- 

 desirable family peculiarities. 



Without caring to dwell on many of my earlier failures to 

 index the finger prints in a satisfactory way, my description 

 shall be confined to that which has proved to be a success. It 

 is based on a small variety of conspicuous differences of pattern 

 in each of many digits, and not upon the numerous minute 

 peculiarities of a single digit. My conclusions are principally 

 based on a study of the ^impressions of all lo digits of 289 

 different persons, but the tables given in the memoir refer only 

 to the first 100 on my list. These are sufficiently numerous to 

 serve as a fair sample of what we might always expect to find, while 

 they are not too cumbrous to print and to discuss in full detail. 



Though I have spoken and shall speak only of impressions, it 

 is not really necessary in forming an index to make any im- 

 pression at all. All the entries that appear in it may be derived 

 directly from the fingers themselves. 



I rely, for the purpose of indexing, on the three elementary 

 divisions of primaries, whorls, and loops. They are severally 

 expressed by the numerals i and 2, 3 and 4, 5 and 6. The 

 reason of this double numeration is that most of the patterns 

 have a definite axis. Those that are formed by ridges which 

 proceed from only one side of the finger, lie in a sloping direc- 

 tion across its axis, the slope being directed according lo the 

 side from which the supply of ridges proceeds. All normal 

 slopes, or those thai are (roughly) parallel to a line drawn from 

 the lip of the forefinger lo the base of the little finger, as well as 

 all the patterns that have no definite axis, are expressed by the 

 odd numerals, i, 3, or 5. All abnormal slopes are expressed 

 by the even numerals 2, 4, or 6. It cannot be too strongly in- 

 sisted that the words right and left are ambiguous, and must not 

 be used here. 



The forefingers are the most variable of all Ihe digits in 

 respect to their patterns, their slopes being almost as frequently 



NO. II 28, VOL. 44] 



abnormal as not ; the third fingers rank next ; the little finger 

 ranks last, as its pattern is a loop in nine cases out of ten. I 

 therefore found it convenient not to index the fingers in their 

 natural order, but so that the sequence of the numerals which 

 express the patterns on the digits should be divided into two 

 groups of three numerals, and two groups of two numerals, as 

 355. 455. 55. 35- The first group 355 referring to the first, 

 second, and third fingers of the left hand ; the second group 

 455 to the first, second, and third fingers of the right hand ; the 

 third group 55 lo the thumb and fourth finger of the left hand ; 

 the fourth group 35 to the thumb and fourth finger of the right 

 hand. The index is arranged in the numerical sequence of 

 these sets of numbers. 



Before translating the patterns into numerals, I find it an 

 excellent plan to draw symbolic pictures of the several patterns in 

 the order in which they appear in the impression, or in the fingers 

 themselves, as the case may be, confining myself to a limited 

 number of symbols [a list of those which have thus far sufficed 

 is given in the memoir ; 5 of them are symmetrical symbols, and 

 9 are tailed and duplicated for the reasons given above, one of 

 each pair being inclined to the right, and the other to the left. 

 The total number of these hieroglyphs is consequently 23]. A 

 little violence has of course to be used now and then, 

 in fitting some unusual pattern to one of these symbols. 

 But we are familiar with such processes in ordinary spelling, 

 where the same letter does duty for different sounds, as a in the 

 words as, ask, ale, and all. The merits of this process are 

 many. It facilitates a leisurely revision of first determinations ; 

 it affords an adequate record of the character of each pattern ; 

 it prevents mistakes between normal and abnormal slopes ; it 

 prevents confusion when changing the sequence of the entries 

 from the order of the impressions to that used in the index ; and, 

 lastly, it affords considerable help to a yet further subdivision of 

 the patterns. 



In making a large and complete index, the symbols would, of 

 course, be cast as movable types, and be printed with the letter- 

 press. 



It appears from the 100 cases that are printed in the memoir 

 that there were 83 different varieties of index numbers when all 10 

 digits are used. Consequently the average number of references 

 required to pick out a single well-defined case from among these 

 100 would be equal to 100 divided by 83 — that is, to about l\. 

 I do not expect from my own reiterated experiences that there 

 would be much trouble due to transitional cases, after a standard 

 collection of doubtful forms had been collected and numbered, 

 so as to insure that different persons should follow a common 

 standard. I find much uniformity in my own judgment. 



Owing to the large effect of correlation, an index based on all 

 the 10 digits is not much superior in efficiency to one that is 

 based on six — namely, upon the first three fingers of both 

 •hands. In the 100 different sets there are, as already said, 

 83 varieties of pattern in the one case, and there are 65 in the 

 other, which roughly accords with the relative efficiency of 5 to 

 4. It is, therefore, a fair question whether it is worth while to 

 impress all the 10 digits. The chief advantage of doing so is to 

 add to the volume of evidence, and to supply data which mutila- 

 tion, or bad scars, or obliteration due to some exceptional cause 

 might render of value. The three fingers of both hands are more 

 than twice as useful for an index as those of one hand only ; 

 again, the three fingers of one hand are nearly twice as useful as 

 two only. I may mention that for my present inquiries into 

 racial and hereditary patterns I am, for various reasons, deal- 

 ing only with the first three fingers of the right hand, and 

 slightly rolling the forefinger, so as to obtain a full impression of 

 its pattern on the side of the thumb. 



When searching through a large number of prints that bear 

 the same index number, in order to find a duplicate of a 

 particular specimen, it is a very expeditious method to fix on 

 some one well-marked characteristic of a minute kind, such as an 

 island, or inclosure, or a couple of adjacent bifurcations, that 

 may present itself in any one of the fingers, and in making the 

 search to use a lens or lenses of low power, fixed at the end of 

 an arm, and to confine the attention solely to looking for that 

 one characteristic. The cards on which the finger marks have 

 been made, may then be passed successively under the lens with 

 great rapidity. 



[It is proposed to exhibit specimens illustrative of this and of 

 the previous memoir, together with appliances for taking 

 impressions from the fingers, at the approaching soiree of the 

 Royal Society.] 



