Bertillon"s anthropometric method insufficient before courts-martial, because possible inaccuracies in measurement, and because of allowable errors.
But ident.i.ty acknowledged following coincident indelible marks, when height, age, and hair fairly correspond.
That is, Bertillon"s collateral evidence is practically primary evidence for such purposes.
There is used for each man an outline figure card giving anterior and posterior surfaces, divided by dotted lines into regions.
These, showing each permanent mark, are filed alphabetically at the Surgeon-General"s office, War Department.
As a man goes out for cause, or deserts, his card is placed in a separate file.
The cards of recruits are compared with the last-mentioned file.
To make this comparison, a register in two volumes is opened, one for light-eyed and one for dark-eyed men. Each is subdivided into a fair number of pages, according to height of entrants, and each page is ruled in columns for body regions. Tattooed and non-tattooed men of similar height and eyes are entered on opposite pages. Recruits without tattoos are not compared with deserters with tattoos; but recruits with tattoos are compared with both cla.s.ses.
On the register S T B M, etc., are used as abbreviations for scar, tattoo, birth-mark, mole, etc.
One inch each side of recorded height allowed for variation or defective measurement.
When probability of ident.i.ty appears, the original card is used for comparison.
Owing to obstacles in inaugurating new system, its practical working began with 1891, and, to include May 1891 [= 5 months, F.G.], out of sixty-two cases of suspected fraud sixty-one proved real.
There was some interesting discussion, both upon this memoir and on a verbal communication concerning the French method, that had been made by M. Jacques Bertillon the statistician, who is a brother of its originator.
It appeared that there was room for doubt whether the anthropometric method had received a fair trial in America, the measurements being made by persons not specially trained, whereas in France the establishments, though small, are thoroughly efficient.
There are almost always moles or birth-marks, serving for identification, on the body of every one, and a record of these is, as already noted, an important though subsidiary part of the Bertillon system. Body-marks are noted in the English registers of criminals, and it is curious how large a proportion of these men are tattooed and scarred. How far the body-marks admit of being usefully charted on the American plan, it is difficult to say, the success of the method being largely dependent on the care with which they are recorded. The number of persons. .h.i.therto dealt with on the American plan appears not to be very large. As observations of this cla.s.s require the person to be undressed, they are unsuitable for popular purposes of identification, but the marks have the merit of serving to identify at all ages, which the measurements of the limbs have not.
It seems strange that no register of this kind, so far as I know, takes account of the teeth. If a man, on being first registered, is deficient in certain teeth, they are sure to be absent when he is examined on a future occasion. He may, and probably will in the meantime, have lost others, but the fact of his being without specified teeth on the first occasion, excludes the possibility of his being afterwards mistaken for a man who still possesses them.
We will now separately summarise the results arrived at, in respect to the two processes that may both be needed in order to effect an identification.
First, as regards _search in an Index_.--Some sets of measures will give trouble, but the greater proportion can apparently be catalogued with so much certainty, that if a second set of measures of any individual be afterwards taken, no tedious search will be needed to hunt out the former set. Including the bodily marks and photographs, let us rate the Bertillon method as able to cope with a register of 20,000 adults of the same s.e.x, with a small and definable, but as yet unknown, average dose of difficulty, which we will call _x_.
A catalogue of 500 sets of finger prints easily fulfils the same conditions. I could lay a fair claim to much more, but am content with this. Now the finger patterns have been shown to be so independent of other conditions that they cannot be notably, if at all, correlated with the bodily measurements or with any other feature, not the slightest trace of any relation between them having yet been found, as will be shown at p. 186, and more fully in Chapter XII. For instance, it would be totally impossible to fail to distinguish between the finger prints of twins, who in other respects appeared exactly alike. Finger prints may therefore be treated without the fear of any sensible error, as varying quite independently of the measures and records in the Bertillon system.
Their inclusion would consequently increase its power fully five-hundred fold. Suppose one moderate dose of difficulty, _x_, is enough for dealing with the measurements, etc., of 20,000 adult persons of the same s.e.x by the Bertillon method, and a similar dose of difficulty with the finger prints of 500 persons, then two such doses could deal with a register of 20,000 500, or 10,000,000.
We now proceed to consider the second and final process, namely, that of identification by _Comparison_. When the data concerning a suspected person are discovered to bear a general likeness to one of those already on the register, and a minute comparison shows their finger prints to agree in all or nearly all particulars, the evidence thereby afforded that they were made by the same person, far transcends in trustworthiness any other evidence that can ordinarily be obtained, and vastly exceeds all that can be derived from any number of ordinary anthropometric data. _By itself it is amply sufficient to convict._ _Bertillonage_ can rarely supply more than grounds for very strong suspicion: the method of finger prints affords certainty. It is easy, however, to understand that so long as the peculiarities of finger prints are not generally understood, a juryman would be cautious in accepting their evidence, but it is to be hoped that attention will now gradually become drawn to their marvellous virtues, and that after their value shall have been established in a few conspicuous cases, it will come to be popularly recognised.
Let us not forget two great and peculiar merits of finger prints; they are self-signatures, free from all possibility of faults in observation or of clerical error; and they apply throughout life.
An abstract of the remarks made by M. Herbette, Director of the Penitentiary Department of the Ministere de l"Interieur, France, at the International Penitentiary Congress at Rome, after the communication by M.
Alphonse Bertillon had been read, may fitly follow.
"Proceeding to a more extended view of the subject and praising the successful efforts of M. Bertillon, M. Herbette pointed out how a verification of the physical personality, and of the ident.i.ty of people of adult age, would fulfil requirements of modern society in an indisputable manner under very varied conditions.
"If it were a question, for instance, of giving to the inhabitants of a country, to the soldiers of an army, or to travellers proceeding to distant lands, notices or personal cards as recognisable signs, enabling them always to prove who they are; if it were a question of completing the obligatory records of civil life by perfectly sure indications, such as would prevent all error, or subst.i.tution of persons; if it were a question of recording the distinctive marks of an individual in doc.u.ments, t.i.tles or contracts, where his ident.i.ty requires to be established for his own interest, for that of third parties, or for that of the State,--there the anthropometric system of identification would find place.
"Should it be a question of a life certificate, of a life a.s.surance, or of a proof of death, or should it be required to certify the ident.i.ty of a person who was insane, severely wounded, or of a dead body that had been partly destroyed, or so disfigured as to be hardly recognisable from a sudden or violent death due to crime, accident, shipwreck, or battle--how great would be the advantage of being able to trace these characters, unchangeable as they are in each individual, infinitely variable as between one individual and another, indelible, at least in part, even in death.
"There is still more cause to be interested in this subject when it is a question of identifying persons who are living at a great distance, and after the lapse of a considerable time, when the physiognomy, the features, and the physical habits may have changed from natural or artificial causes, and to be able to identify them without taking a journey and without cost, by the simple exchange of a few lines or figures that may be sent from one country or continent to another, so as to give information in America as to who any particular man is, who has just arrived from France, and to certify whether a certain traveller found in Rome is the same person who was measured in Stockholm ten years before.
"In one word, to fix the human personality, to give to each human being an ident.i.ty, an individuality that can be depended upon with certainty, lasting, unchangeable, always recognisable and easily adduced, this appears to be in the largest sense the aim of the new method.
"Consequently, it may be said that the extent of the problem, as well as the importance of its solution, far exceeds the limits of penitentiary work and the interest, which is however by no means inconsiderable, that penal action has excited amongst various nations.
These are the motives for giving to the labours of M. Bertillon and to their practical utilisation the publicity they merit."
These full and clear remarks seem even more applicable to the method of finger prints than to that of anthropometry.
CHAPTER XI
HEREDITY
Some of those who have written on finger marks affirm that they are transmissible by descent, others a.s.sert the direct contrary, but no inquiry hitherto appears to justify a definite conclusion.
Chapter VIII. shows a close correlation to exist between the patterns on the several fingers of the same person. Hence we are justified in a.s.suming that the patterns are partly dependent on const.i.tutional causes, in which case it would indeed be strange if the general law of heredity failed in this particular case.
After examining many prints, the frequency with which some peculiar pattern was found to characterise members of the same family convinced me of the reality of an hereditary tendency. The question was how to submit the belief to numerical tests; particular kinships had to be selected, and methods of discussion devised.
It must here be borne in mind that "Heredity" implies more than its original meaning of a relationship between parent and child. It includes that which connects children of the same parents, and which I have shown (_Natural Inheritance_) to be just twice as close in the case of stature as that which connects a child and either of its two parents. Moreover, the closeness of the fraternal and the filial relations are to a great extent interdependent, for in any population whose faculties remain _statistically_ the same during successive generations, it has been shown that a simple algebraical equation must exist, that connects together the three elements of Filial Relation, Fraternal Relation, and Regression, by which a knowledge of any two of them determines the value of the third. So far as Regression may be treated as being constant in value, the Filial and the Fraternal relations become reciprocally connected. It is not possible briefly to give an adequate explanation of all this now, or to show how strictly observations were found to confirm the theory; this has been fully done in _Natural Inheritance_, and the conclusions will here be a.s.sumed.
The fraternal relation, besides disclosing more readily than other kinships the existence or non-existence of heredity, is at the same time more convenient, because it is easier to obtain examples of brothers and sisters alone, than with the addition of their father and mother. The resemblance between those who are twins is also an especially significant branch of the fraternal relationship. The word "fraternities" will be used to include the children of both s.e.xes who are born of the same parents; it being impossible to name the familiar kinship in question either in English, French, Latin, or Greek, without circ.u.mlocution or using an incorrect word, thus affording a striking example of the way in which abstract thought outruns language, and its expression is hampered by the inadequacy of language. In this dilemma I prefer to fall upon the second horn, that of incorrectness of phraseology, subject to the foregoing explanation and definition.
The first preliminary experiments were made with the help of the Arch-Loop-Whorl cla.s.sification, on the same principle as that already described and utilised in Chapter VIII., with the following addition. Each of the two members of any couplet of fingers has a distinctive name--for instance, the couplet may consist of a finger and a thumb: or again, if it should consist of two fore-fingers, one will be a right fore-finger and the other a left one, but the two brothers in a couplet of brothers rank equally as such. The plan was therefore adopted of "ear-marking" the prints of the first of the two brothers that happened to come to hand, with an A, and that of the second brother with a B; and so reducing the questions to the shape:--How often does the pattern on the finger of a B brother agree with that on the corresponding finger of an A brother? How often would it occur between two persons who had no family likeness? How often would it correspond if the kinship between A and B were as close as it is possible to conceive? Or transposing the questions, and using the same words as in Chapter VIII., what is the relative frequency of (1) Random occurrences, (2) Observed occurrences, (3) Utmost possibilities?
It was shown in that chapter how to find the value of (2) upon a centesimal scale in which "Randoms" ranked as 0 and "Utmost possibilities" as 100.
The method there used of calculating the frequency of the "Random" events will be accepted without hesitation by all who are acquainted with the theory and the practice of problems of probability. Still, it is as well to occasionally submit calculation to test. The following example was sent to me for that purpose by a friend who, not being mathematically minded, had demurred somewhat to the possibility of utilising the calculated "Randoms."
The prints of 101 (by mistake for 100) couplets of prints of the right fore-fingers of school children were taken by him from a large collection, the two members, A and B, being picked out at random and formed into a couplet. It was found that among the A children there were 22 arches, 50 loops, and 29 whorls, and among the B children 25, 34, and 42 respectively, as is shown by the _italic_ numerals in the last column, and again in the bottom row of Table XX. The remainder of the table shows the number of times in which an arch, loop, or whorl of an A child was a.s.sociated with an arch, loop, or whorl of a B child.
TABLE XX.
_Observed Random Couplets._
+----------------------------------------------------+A children.Totals inB children.------------------------B children.Arches.Loops.Whorls.------------------------------------------------Arches5128_25_Loops8188_34_Whorls92013_42_------------------------------------------------Totals in A}children }_22__50__29_101+----------------------------------------------------+
TABLE XXI.
_Calculated Random Couplets._
+----------------------------------------------------+A children.Totals inB children.------------------------B children.Arches.Loops.Whorls.------------------------------------------------Arches5001250725_25_Loops6801700986_34_Whorls84021001218_42_------------------------------------------------Totals in A}children }_22__50__29_101+----------------------------------------------------+
The question, then, was how far calculations from the above data would correspond with the contents of Table XX. The answer is that it does so admirably. Multiply each of the italicised A totals into each of the italicised B totals, and after dividing each result by 101, enter it in the square at which the column that has the A total at its base, is intersected by the row that has the B total at its side. We thus obtain Table XXI.
We will now discuss in order the following relationships: the Fraternal, first in the ordinary sense, and then in the special case of twins of the same set; Filial, in the special case in which both parents have the same particular pattern on the same finger; lastly, the relative influence of the father and mother in transmitting their patterns.