Alphonse Bertillon and

the measure of man
More expert than Sherlock Holmes

“Every measurement slowly reveals the workings of the criminal. Careful observation and patience will reveal the
truth.” Thus wrote the French criminologist Alphonse Bertillon, who died 100 years ago in the spring of 1914. Almost
forgotten today, he was the most famous criminologist of his time, a household name for his ingenious method of
identifying criminals by carefully measuring 11 key dimensions of their bodies. He did more. He established the
standardised procedures used in police forces to this day. Richard Farebrother and Julian Champkin look at a
flawed near-genius.

“I came to you, Mr. Holmes, because I Readers of the detective stories of Sir Arthur Co- procedural” is a modern form of detective novel.
recognized that I am myself an unprac- nan Doyle and Dame Agatha Christie have little Bertillon set the procedures.
tical man and because I am suddenly excuse for not having heard of Alphonse Bertil- There is more. He had a hand in the creation
confronted with a most serious and ex- lon. Other thriller writers also mention him, or of a rather significant statistical concept as well.
traordinary problem. Recognizing, as I do, his method of identifying criminals by measuring Bertillon pre-dated fingerprinting – though not
that you are the second highest expert in what we might as well call their vital statistics. by very long. (His relationship with the concept
Europe –” He called it anthropometry, or “people measur- of fingerprints was complicated, as we shall see.)
“Indeed, sir! May I inquire who has ing”; most people just called it the Bertillon Sir Francis Galton, the champion and promoter of
the honour to be the first?” asked Holmes system. Holmes and Agatha Christie’s detectives fingerprinting as well as the begetter of a large
with some asperity. are of course fictional; but Alphonse Bertillon chunk of modern statistics, hugely admired Ber-
“To the man of precisely scientific mind was very much a real-life policeman. And in his tillon, visited him in his Paris police headquar-
the work of Monsieur Bertillon must al- day he was almost as famous as Sherlock Holmes. ters, and had his mug shot and anthropometric
ways appeal strongly.” He was rather more than the deviser of an measurements taken by the great man himself
“Then had you not better consult him?” ingenious way of measuring people for the – Bertillon’s photos of Galton are reproduced in
“I said, sir, to the precisely scientific police. Photography of suspects, in uniform for- the September 2011 issue of Significance. The
mind. But as a practical man of affairs it mats and poses – full-face and profile – played men shared many characteristics. Both were
is acknowledged that you stand alone. I a role – albeit a secondary one – in his system: obsessive, both were passionate about measur-
trust, sir, that I have not inadvertently –” he invented the mug shot. The chair and camera ing everything that could be measured and
“Just a little,” said Holmes. he took them with were bolted to the floor to some things that could not really be measured
ensure standard distances and positions. System besides. Galton was eccentric to the point of
(Sir Arthur Conan Doyle, and standardisation lay at the heart of every-
The Hound of the Baskervilles, 1902) thing he did. He brought in standardised police
records – and, just as important, systematic ways
of classifying and searching those records: he
“You saw that note I wrote just now?” said was a pioneer of data mining. He introduced the
the French detective. “It was to my peo- recording of crime scenes by photography. The
ple at the inn. Yesterday I received from cliché of the chalk outline of the body on the
France the fingerprints and the Bertillon carpet owes much to him, though there is no
measurements of the so-called Captain record of his actually using it himself. Instead,
O’Neill. In a few moments we shall know he mounted cameras on high tripods over the
whether you are the man!” murder victim to take aerial views from a stand-
ardised height. It is not too much to say that
(Agatha Christie, modern scientific – or statistical – investiga- Self portrait of Alphonse Bertillon, dated August 22nd
The Secret of Chimneys, 1925) tion of crime began with Bertillon. The “police 1900

36 april2014 © 2014 The Royal Statistical Society

 measurements somewhere along it. Bertillon
concluded that no two people have exactly the
same physical dimensions. And if you chose the
right dimensions to measure, to the right degree
of accuracy, you would end up with a set of num-
bers that were unique to that individual, that
would remain with him for life – and that would
infallibly identify him to any investigator at any
later date.
He visited local prisons to measure the
heights and lengths of limbs of inmates. He
looked at the shapes and angles of their noses,
the distance between their eyes. Using specially
Illustration from The Speaking Portrait, an article from Pearson’s Magazine, 1901, illustrating the principles of
devised callipers, he measured their heads in
Alphonse Bertillon’s anthropometry all kinds of directions. The prisoners laughed at
him. So did the warders.
The system he wanted should be quantitative,
near-insanity, Bertillon was socially gauche, bad- full advantage. If a thief or a pickpocket was based on accurate, repeatable measurement. It
tempered, obstinate to an extraordinary degree, arrested, the usual practice was simply to give should be systematic, objective and, he was to
and was accused by superiors – and indeed by a false name (Dupont, it seems, was the name claim, infallible. The system he came up with
the President of France – of being insane. These they mostly chose – police records of the time took 11 measurements from each individual. No
two near-insanities in combination brought use- are unnaturally full of them) and claim it was other person in the world would share that set
ful results. Galton looked at two components of the first time he or she had been in trouble. The of measurements.
Bertillon’s measurement system – the length of only way to disprove it was though the memory From the body he took the height, the width
the forearm and the height. He considered that of police officers. Some, indeed, had phenomenal of the outstretched arms, and the sitting height.
they were related – most tall people have long memories: Eugène Vidocq, the legendary former He took the length of the head, its breadth,
arms – but that one did not cause the other. Thus criminal and first head of the French police, was the distance between the cheek bones and the
Galton was led in 1888 to the statistical concept said never to have forgotten a face. That was length of the right ear. The length of the left
of correlation1. the exception. Ordinary police constables were foot, the left middle and little fingers, and of
Bertillon certainly came of sound statistical paid 5 francs for each suspect they recognised as the left arm from the elbow to the tip of the
stock. His maternal grandfather, one Achille having been arrested before. Rumour had it that outstretched middle finger completed the total.
Guillard, was a pioneer statistician who in 1855 they would split the sum with any criminal who He preferred making measurements on the left
wrote one of the first books on demography, confessed to having form. side as it was less influenced by the effect of
Eléments de statistique humaine ou démographie The descriptions on the cards that young work. (Right-handed manual workers develop
comparée, and, indeed, coined the word. His fa- Alphonse was so laboriously copying were vague, larger right arms.)
ther, Louis-Adolphe Bertillon, was a statistician written in whatever terms the constable respon- Bertillon – with statistics in his blood – cal-
at the school of anthropology in Paris; his older sible thought fit. And when he had filled them culated that the chance against all 11 points
brother, Jacques, became a medical statistician in, the cards were unsearchable: if you were being repeated in any two individuals was
of note who discovered the correlation between seeking to identify a redhead with a missing 268 435 456 to 1. The population of France was
suicide rates and divorces, claiming that both forefinger who claimed his name was Dupont under 70 million, so that should have provided
were associated with societies in a state of in- you would have to look though all the Duponts, some certainty; but to make assurance doubly
stability. Young Alphonse was the black sheep and every other name as well – every card in the sure he added three more points, qualitative
of the family. Twice expelled from school – once entire index – so you stood effectively no chance rather than quantitative: the colours of the eyes,
from a school for disruptive children – and set- of finding him. Unsystematic, unsearchable: the hair and skin. A separate section of the record
ting fire to his desk in a third, he had a mediocre cards were a waste of time. card set down tattoos, scars and the like.
army career (perhaps because he was so tone Bertillon objected to wasting his time. But No less ingenious was his method of making
deaf that he could only recognise bugle calls by he dared not throw up his job, for fear of his his database searchable. His cards were not filed
counting the notes), then failed as a teacher. father. Here his passion for order came into play by alphabetical order of names. Instead they
Only by pulling strings was his father eventually – as did his knowledge of his father’s statisti- were divided into three groups, of those with
able to get him, aged 26, a job as a lowly clerk in cal work: Bertillon père had laboured to prove small, medium or large lengths of head – which
the Paris police, copying out, in a freezing cellar- that each human being had unique variations in he called alpha, beta and gamma. Each group
basement, the identification forms that had to be physical characteristics. A second, more famous, was again divided in three, this time by breadth
filled out for each prisoner. As Alphonse rapidly statistical source also inspired him. The Belgian of head – again alpha, beta and gamma. The next
realised, these forms were essentially useless. Adolphe Quetelet had produced a study of French triple subdivision was by cheek distance, and so
This was the Paris of 1879. Branding of crime statistics; he had also, in his famous work on down. Searching this database was then easy.
convicted criminals – a useful way of spotting of 1835, Sur l’homme et le développement de To see if the man in front of him had been
an ex-con and possible malefactor – had been ses facultés, ou: Essai de physique sociale, intro- arrested before, the police officer took the
banned fifty years before; fingerprinting was duced the concept of the “average man”. Bodily suspect’s Bertillon measurements. If the head
yet to be recognised; photography was in its measurements – height, weight and the like – in was long and broad he looked first at the “long
infancy. There was, therefore, simply no way of a population, said Quetelet, were distributed head” group of file cards, and, within them, at
identifying an habitual offender. Criminals took along a normal curve, with each individual’s the “broad head” cards; within that group, by

april2014 37
now nine times smaller, he took out the section
for the appropriate cheek width, and so on. If he
got down to a card that matched all 11 of the
suspect’s measurements the identification was,
supposedly, certain.
Bertillon wrote to his superior, who ignored
him. He wrote again, quoting Quetelet. This time
his boss replied – to Bertillon’s father, warning
him that any more such displays of madness from
the son would result in the loss of his job. His
father advised Bertillon to wait before trying
again, on the grounds that bosses, useless or
competent, do not last for ever.
And indeed, the obstructive superior was re-
placed by one who, mindful of the father’s influ-
ence, gave him three months to identify a repeat
offender. Bertillon set to work, measuring prison
inmates on the verge of release, filling in their
11-point data cards like mad, assisted by a young
woman he had met crossing a road and whom he
had asked to help him mainly on the strength
of her small and neat handwriting (those cards
needed legibility). She filled out 7336 cards; then
he married her. (For these and other details we are
indebted to what seems to be the only biography A class studying the Bertillon method of criminal identification, c. 1910–1915. Source: Library of Congress
of Bertillon in English2. By Dorothy and Thomas
Hoobler, it is contained in what is actually the measured just on a vague suspicion. The police paramount in police work, he himself preferred
story of the theft of the Mona Lisa.) And, after covertly arranged a society function of the type obstinacy, even – or especially – when he was
two of the allotted three months had passed, the that Vernet would attend; when the guests were wrong. A drawback to his system, frequently com-
first match turned up. One of those released and assembled a surprise entertainment was an- plained of by hapless police constables, was the
measured ex-cons had offended again – and had nounced – a demonstration of the method of the time and skill needed to take the measurements.
been re-measured and identified by Bertillon. It famous Monsieur Bertillon using members of the He refused to believe it. “Anyone who is not an
was not, to be sure, a very major crime: the man audience as his guinea-pigs. The suspect moved imbecile could learn to measure in five minutes,
had stolen an empty bottle. But it was a start. to the door, but, publicly hailed from the stage, and never forget the process.” This was not actu-
The method worked. By the end of the year he was unable to avoid being “volunteered” and ally true. Variation in the tightness with which
had identified 49 repeat offenders. His method measured. The Bertillon measurements matched; callipers were applied to the body frequently gave
was officially adopted. It identified 800 suspects Vernet/Simon hanged himself in jail rather than different measurements for the same individual.
in its first three years, and 3500 in its first ten. face another stretch in Cayenne. His attitude to fingerprints was another example:
The database grew and became huge; and it had By 1888 Bertillon was director of a specially despite that friendship with Galton, and their
some great successes. created Bureau of Identification, wildly praised growing use worldwide, he refused to consider
It was a time when bomb plots by anarchists at home and with a worldwide reputation. His them as anything more than a minor adjunct to
were greatly feared. Bertillonage captured the system had been adopted by police forces eve- his system. He did add them to his record cards.
most famous of them, one Ravachol, who had set rywhere. The French government had decreed But when the Argentine detective Juan Vucetich,
off two bombs and a wave of panic in the city. that vagrants and those of no fixed abode should who had pioneered their use in practice, came
Ravachol was guillotined, Bertillon was made a carry cards with their Bertillon measurements to to Paris to visit him, Bertillon kept him waiting
member of the Legion of Honour for catching identify them: ID cards are another innovation for hours, then finally opened the office door,
him, and became a public hero into the bargain. we owe to Bertillon. shouted “You have done me incalculable harm”
Another extraordinary identification was of a He was, however, far from unflawed. Though – and slammed the door in his face.
swindler by the name of Simon. He was, official- proclaiming that science and evidence should be Bertillon displayed all his arrogance, obsti-
ly, dead. He had been tried, convicted and sent nacy and wrong-headedness – as well as more
to the penal colony of Cayenne, on the South Bertillonage is the greatest and most brilliant unpleasant characteristics and being flagrantly
American coast. He and a fellow-prisoner stole invention the nineteenth century has wrong besides – in the Dreyfus case, the cause
a canoe and tried to escape. It was assumed produced in the field of criminology. Thanks célèbre which rocked France for a decade and all
they never made it. A body was found, with the to a French genius, errors of identification but brought down governments. Some military
face smashed beyond recognition but wearing will soon cease to exist not only in France papers were leaked to the German embassy. Cap-
Simon’s clothes. A few years later a successful fi- but also in the entire world. Hence judicial tain Alfred Dreyfus was tried and convicted, ap-
nancier named Vernet appeared in Paris, wealthy errors based on false identification will parently simply because he was Jewish, and sent
and well connected. The head of the Paris police likewise disappear. Long live Bertillonage! to Devil’s Island. Evidence came to light pointing
believed him, on nothing more than a hunch, Long live Bertillon! to a Major Esterhazy as the true culprit, but the
to be Simon, but the suspect was moving in (Contemporary praise of Bertillon) military authorities suppressed it. Newspaper
high social circles – too high to be detained and articles, campaigns, courts-martial, retrials, and

38 april2014
libel cases split French society – and Bertillon phantasies. He was at last taking his revenge for He refused to go along with fingerprinting
appeared in almost all of the legal actions loudly all the insults of those who had called him a fit as well. “My measurements are surer than any
proclaiming Dreyfus’s guilt. An early trial centred subject for an asylum of the insane.” fingerprint pattern”, he said. He was wrong. Two
on a forged document. Bertillon pronounced His attitude split the hitherto solidly liberal remarkable cases proved it, one in Britain, one in
that it was undoubtedly in Dreyfus’s handwrit- family. His brother Jacques, who was married to America. The British one involved identical twins,
ing, disguised first using a square grid to form a Jew, did not speak to him for years. the Fox brothers, confusingly named by their
the letters, then disguised again in a supposed Eventually three of France’s most eminent doubtless loving mother Ebenezer Alfred Fox and
double bluff by trying to imitate an imitation scientists re-examined the disputed handwriting, Alfred Ebenezer Fox. Both were criminals, and
of his own handwriting. To which one can only using more sophisticated measuring instruments they kept giving alibis to each other. Their Bertil-
try to push one’s jaw back up. He stuck to this and coming up with a statistical probability for lon measurements were exactly the same, which
convoluted argument, with still more convoluted their answer. It comprehensively demolished should have caused little surprise, and they
diagrams, through a series of subsequent trials. Bertillon’s argument. He never accepted it. On could be told apart only by their fingerprints.
It was when the Minster for War introduced his his deathbed the government offered him the More remarkable was the American case of Will
star anti-Dreyfus witness to the President of the additional award of the rosette to his Legion of West, arrested in Kansas in 1903 and measured in
Republic that the aforesaid President described Honour, if only he would acknowledge his error Leavenworth Gaol. His measurements showed him
Bertillon as having escaped from an asylum. about the handwriting of Dreyfus, now long since to be one William West, arrested for murder two
It got worse. The novelist Emile Zola, who pardoned. Bertillon shouted “No! Never! Never!” years before. Unfortunately William West was still
exposed the scandal in his famous open letter, They were among the last words he spoke. serving his sentence, and, by bizarre coincidence,
J’Accuse...!, was prosecuted for libel. Again But to regard him as reactionary in every- in the same prison. The two men, though appar-
Bertillon took the stand – and was reduced thing would be too simplistic. The theories of ently not related, were identical.
by Zola’s counsel first to incoherence, then to the Italian Cesare Lombroso were popular at the Today, of course, his system has been super-
helpless silence. And when Dreyfus was brought time: that there were “criminal types”, born that seded by fingerprinting, and, more recently, by
back from Devil’s Island to face a second court- way, who could be identified from their appear- DNA analysis. Fingerprints had two great advan-
martial, Bertillon testified once more. An English ance – low forehead, protruding brows, ape-like tages over Bertillon’s system. It was far quicker
journalist described his performance: “Now and face. Bertillon’s (by now vast) database of meas- and easier to take a suspect’s fingerprints than
again M. Bertillon’s voice rose in hateful shrieks. urements and photographs of criminals could to go through the complicated measurement
There were interludes when he clenched his fist have been easily cherry-picked to foster such procedure of the Bertillon system; and, unlike
and struck the bar, swearing that Dreyfus was beliefs. (Fifty years later the Nazis used very Bertillon’s measurements, fingerprints leave their
the traitor. The voice rang out with passion and similar techniques to “identify” supposed Jewish own traces at the scene of the crime. It made
excitement. You beheld in him the man vain or degenerate ancestry.) But Bertillon refused to no difference to Bertillon. Thanks to him, France
unto madness with confidence in his atrocious go along (see box). was the last country to adopt fingerprinting.
Is Bertillon’s an outdated system of historical
interest only? Perhaps, but today facial recogni-
Bertillon and the criminal type tion algorithms can be applied to images from
Bertillon was interviewed in 1894 by the American journalist Ida Tarbell, for McClure’s magazine. security cameras, at airports and in streets, to
identify wanted individuals. The algorithms
“But your archives, M. Bertillon?” I asked. “Are you not going to use your observations work by taking distances between key points of
for scientific deductions, for anthropological conclusions, as, for instance, to establish a the face and comparing them statistically to a
criminal type?” known database of images. They use many more
“Undoubtedly,” he responded, “the statistics of the service will be used more and more measurements than Bertillon managed – dozens
for ethnographical and anthropological statistics. I have already done something with or hundreds for the face alone – but the principle
them. Here is a chart showing the color of the eyes in the different parts of France, from is much the same. Perhaps he was simply ahead
the maroon of the Spanish border to the blue of the Channel; and there is another, giving of his time.
the relative length and breadth of the head. As for the criminal type, that is a delicate
question.” References
“Then you have never sought to confirm the doctrine of Lombroso’s school, that certain 1. Stigler, S. M. (1999) Statistics on the
anatomical characteristics indicate the criminal?” Table: The History of Statistical Concepts and Methods.
“No; I do not feel convinced that it is the lack of symmetry in the visage, or the size of Cambridge, MA: Harvard University Press.
the orbit, or the shape of the jaw, which make a man an evil-doer. A certain characteristic 2. Hoobler, D. and Hoobler, T. (2009) The
may incapacitate him for fulfilling his duties, thus thrusting him down in the struggle for Crimes of Paris: A True Story of Murder, Theft, and
life, and he becomes a criminal because he is down. Lombroso, for example, might say Detection. New York: Little, Brown and Co.
that, since there is a spot on the eye of the majority of criminals, therefore the spot on
the eye indicates a tendency to crime; not at all. The spot is a sign of defective vision, Richard William Farebrother was a member of the
and the man who does not see well is a poorer workman than he who has a strong, keen teaching staff of the University of Manchester from
eyesight. He falls behind in his trade, loses heart, takes to bad ways, and turns up in 1970 until 1993 when he took early retirement
the criminal ranks. It was not the spot on his eye which made him a criminal; it only on medical grounds. From 1993 until 2001 he was
prevented his having an equal chance with his comrades. The same thing is true of other Honorary Reader at the same university.
so-called criminal signs. One needs to exercise great discretion in making anthropological He has published three books and more than 150
deductions. Nevertheless, there is no doubt but that our archives have much to tell on all research papers on subjects including econometric
questions of criminal anthropology.” theory, computer algorithms, statistical distributions,
statistical inference and the history of statistics

april2014 39
Copyright of Significance is the property of Wiley-Blackwell and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.