by Anne Wingate
Interpol, of course, is able to work with both the Vucetich and the Henry systems; however, it is not an actual police agency but rather an international clearinghouse of police information. It has no jurisdiction or arrest powers of its own.
And how did fingerprinting reach the United States? In 1902, Dr. H.P. de Forest was able to require its use in the New York State Civil Service Commission, to cut down on fraud in the taking of tests. In 1904, during a world's fair in St. Louis, Missouri, a Scotland Yard inspector taught fingerprinting to several police officers in the United States. The entire U.S. armed forces had taken up fingerprinting by 1908.
The Will West Case
The value of the fingerprint system was conclusively proven in Leavenworth Federal Penitentiary in 1903. A prisoner was brought in by the name of Will West. Admission personnel, photographing and measuring him, insisted he had been there before, but he insisted he had not, and his conviction listed him as a first offender. His Bertillon measurements (see sidebar) and photographs seemed identical to those already on record for William West. But on further
checking, personnel found that this could not be William West — because William West was still in prison!
When the two men were placed side by side, they looked like identical twins. But despite theories expressed even today that they were related—fraternal or possibly even identical twins — no amount of investigation has ever been able to discover even a distant relationship, and their fingerprints were not at all alike. (This is possibly significant in terms of the theory that they were related. Identical twins, even nonidentical siblings, often have similar fingerprints and may even have the same fingerprint classification. But identical fingerprint classifications do not mean that the prints are identical.)
What Are Bertillon Measurements?
French file clerk Alphonse Bertillon invented, in 1879, the first really successful means of criminal identification. Prior to that time, the only means of recognizing a criminal who was trying to get by with an assumed name was by the good memory of police officers. The few descriptions that were made were so vague as to be useless: "tattoo on arm" meant nothing unless a description of what the tattoo looked like and where the tattoo was located on which arm was also provided.
Bertillon's system, called Bertillonage, included a set of fourteen different measurements of different parts of the body, coupled with a portrait parle, a very precise system of careful description. Bertillon, however, did an atrocious job of presenting his idea, and when officials finally agreed to let him try it, he was told that if he did not succeed in catching one criminal in the act of changing his name within three months, the entire system would be scrapped.
The chances of having someone arrested, released and then re-arrested under a different name within three months seemed slim; however, Bertillon set to work-and succeeded. Bertillonage at once was the rage worldwide. However, it had several major problems: (1) its success depended on making careful and time-consuming measurements of every person arrested, and at that time very few police were either willing or able to take that time; (2) the portrait parle system, although a great improvement over previous systems of description, was too difficult for most people
to master; and (3) it was impossible to use Bertillonage on a person who was not already in hand.
Bertillon was extremely bitter as he watched his system go down in defeat before the science of fingerprints. As long as he lived, he fought to keep it alive, at least in France; but immediately on his death, Bertillonage was abandoned even there.
The Roscoe Pitts Case
People have attempted to beat the fingerprint system. In 1941, Robert J. "Roscoe" Pitts was released from Alcatraz, determined to beat the fingerprint system next time. He robbed a warehouse and service station in North Carolina in May, and then headed to New Jersey, where plastic surgeon Leopold Brandenburg agreed to replace the skin of his fingertips with skin grafts from his sides. When the lengthy and painful surgery, which involved having Pitts's hands temporarily grafted to his sides, was finally over, Pitts's fingertips were completely smooth. But in October, when he was arrested in El Paso, Texas, police had no trouble at all identifying him by the prints of his fingers below the tips. Furthermore, a Treasury agent who saw Pitts twenty years later told me that even then Pitts was unable to use his fingertips normally because of their extreme sensitivity to pressure and heat.
With that debacle, petty hoodlum Roscoe Pitts found his way into history. If he hadn't tried to get rid of his fingerprints, nobody would ever have heard of him.
Are Criminals too Smart?
I have heard people—even police officers—say that there has been so much publicity on fingerprints that criminals aren't dumb enough to leave them anymore. Fortunately, that is absolutely untrue. U.S. Department of Justice figures indicate that a constant minimum of one-third of all property crimes result in usable fingerprints, if police would just look for them. The sad fact, however, is that too many police have stopped looking for fingerprints, with the result that too many criminals are getting away with crimes they should not be getting away with.
When my house in Fort Worth was burglarized, it took me several telephone calls to get ident officers to come to the scene. When one finally showed up, she informed me in superior tones that it is impossible to get fingerprints from aluminum screen frames.
That was a lie; aluminum screen frames are one of the best surfaces. But I pretended to assume she was simply ill-informed; I explained my background, borrowed her fingerprint equipment, and lifted several beautiful prints, which I turned over to her. To the best of my knowledge, nothing was ever done with the prints—even when a couple of teenagers who had been burglarizing houses in my neighborhood were arrested.
I visited the Fort Worth ident office and learned why. Despite banks of stored fingerprint cards, they had one person working apparently only part-time on latent comparison, and apparently no attempt at nonsuspect identification was even being made.
Introducing AFIS
In real life, every burglary, robbery, recovered stolen auto, theft from auto, sexual assault by an unknown assailant, forgery, and counterfeiting—as well as any other crime in which the use of human hands was involved—should be followed, as soon as possible, by a diligent search for fingerprints. If this were done, and if all legible lifted fingerprints were then checked through AFIS—the Automated Fingerprint Identification System—the conviction rates in some areas would more than triple. In fiction, you may decide whether the department you're writing about does this right, does it in what police would call a half-assed manner, or does it not at all. If you're writing about a real department, find out how good a clearance rate (that is, what percentage of reported crimes are solved) the department has and how much use it makes of fingerprints. Ask politely; people tend to get defensive about these things. In writing about the Fort Worth Police Department, for fairly obvious reasons I tend to avoid mentioning fingerprints. My other series is set in a fictitious town, but it's near the real town of Galveston, and Galveston has a terrific ident section, which I mention fairly often.
Suspect Ident, Nonsuspect Ident
What is the difference between a suspect ident and a nonsuspect ident?
In a suspect ident, a suspect has been developed by other means, and all the identification technician has to do is to compare the latent with one fingerprint card. Anyone who cannot make an ident in that way has no right to call himself or herself an identification technician.
In a nonsuspect ident situation, there is no evidence except the lifted latent. The identification technician must search the latent—that is, compare the latent to a set of fingerprint cards. In Albany, Georgia, we maintained separate sex-crime files and burglar files, so that the print would be searched through the appropriate files (a robber would usually be found in the burglar file, as few people graduate to robbery without first trying burglary). Only in extraordinary cases would we search the entire files; comparing
a single print to 10,000 or more fingerprint cards by hand is a monumental task.
Prior to AFIS, even a large police department identification section making as many nonsuspect idents a month as Doc and I were would have had the right to preen. But now, with AFIS, a police department of any size at all that is not making at least twenty to thirty idents a month must be considered lazy or ignorant. One identification officer proudly told me that his department had about a 15 percent ident rate; that is, they made idents on about 15 percent of all latents lifted. That sounded pretty good to me—until I found out his department, in a town eight times the size of Albany, was making fewer latent lifts than Doc and I had been making idents.
How AFIS Works
Let's step back for a moment and talk about fingerprints themselves.
And why do I want to do this? Because a mystery writer can be made to look extremely stupid if s/he misdescribes fingerprints. My dear friend Elizabeth Linington/Lesley Egan/Dell Shannon once severely misdescribed a fingerprint in one of her novels. I gave her a book about fingerprints. In her next novel, she described a "tented whorl." There is no such thing.
Fingerprint Types
There are three main types of fingerprints: arches, loops and whorls. About 35 percent of all fingerprints are whorls; about 5 percent of all fingerprints are arches; and about 60 percent of all fingerprints are loops. These are subdivided into plain arches and tented
arches; radial loops and ulnar loops; and plain whorls, double loop whorls, central pocket loop whorls, and accidentals. People often have different patterns on different fingers; whorls are most often found on the thumbs, and arches are most often found on the index and middle finger. Now let's have a look at some of these patterns.
Figure 5-1 illustrates the eight different fingerprint patterns. The drawing of the radial and ulnar loops assumes that the prints are from the right hand; in fact, a radial loop is one in which the direction of flow tends to be in the direction of the thumb, and an ulnar loop is one in which the direction of flow tends to be in the direction of the little finger. Ulnar loops are far more common than radial loops, and radial loops tend most commonly to be found only on the index or index and middle fingers.
There are 1,024 possible primary classifications of fingerprints—and where do we get the primary classification? Figure 5-2 is the rolled print portion of a fingerprint card; remember that below it will be the plain—nonrolled—prints, with all four fingers on each hand printed together, the left hand first, then the left thumb, then the right thumb, then the right four fingers.
The top number in this chart is the finger number. All fingerprint cards use the same finger numbers, so that the prints are always taken in the same order and so that an identification technician will always know, for example, that the number 1 finger is the right thumb and the number 10 finger is the left little finger. Each numbered finger is given a numerical value as shown above, but only the whorls are counted. Arches, tented arches and loops are called "nonnumbered" patterns. The primary classification appears as a fraction, with the sum of the first number in each set of two—that is, fingers number 1, 3, 5, 7 and 9—plus a 1 which is added to each set of numbers, as the denominator and the sum of the second set of numbers—fingers 2, 4, 6, 8 and 10—plus the added 1 as the numerator. For example, suppose that we have whorls on both thumbs and loops everywhere else (this is an extremely common pattern). The primary classification would have a denominator of 16 (from the right thumb, finger #1) plus 1 (the additional 1), or 17, and a numerator of 4 (from the left thumb, finger #6) plus 1 (the additional 1) of 5 —hence, 5/17. The single most common primary classification, taking in over half of all fingerprint patterns, is 1/1 — that is, there are no whorls at all; therefore, none of the numerical values count except the additional 1 which is always added.
You don't have to understand this. You don't even have to try. I have explained it in far greater detail than this to people who then stare at me and exclaim, "That's impossible."
The entire FBI fingerprint classification system—a major refinement of the Henry system—with all its primaries, secondaries, subsecondaries, sub-subsecondaries, finals and so forth, is extremely complicated; even some fingerprint technicians do not understand it fully, and there is no earthly reason to give it here. If you really want to learn it, get a copy of the FBI fingerprint handbook The Science of Fingerprints, available from the United States Government Printing Office for about eight dollars; several other very fine fingerprint books are listed in the bibliography.
The NCIC System
The NCIC fingerprint system, although it is based on the Henry system, is not quite the same. We need to get to a little more information before explaining it.
Figure 5-3 shows a loop pattern, with arrows pointing to the delta and the core. A loop is measured by counting the ridges intervening between the delta and the core.
Figure 5-4 shows a whorl pattern. Whorls theoretically always have at least two deltas (only the very rare accidental whorl can have more than two deltas); in fact, as most fingerprint people know, some whorls do not have two deltas. But a whorl is measured by tracing the ridge from the left delta. If it goes more than two ridges inside the right delta, as this one does, it is an inner trace whorl. If it meets the right delta within two ridges one way or the other, it is a meeting trace whorl. If it goes more than two ridges outside the right delta, it is an outer trace whorl. Arches and tented arches are not measured.
For an NCIC classification, a fingerprint would be listed as follows:
A plain whorl would be shown as P plus the tracing. Thus, an inner trace plain whorl would be listed as PI.
A central pocket loop whorl would be listed as C plus the tracing. Thus a central pocket loop whorl with a meeting trace (which is very uncommon—a central pocket loop is most often an inner or outer trace) would be listed as CM.
A double loop whorl would be listed as D plus the tracing. Thus a double loop whorl with an outer trace would be listed as DO.
An accidental—which effectively is any print that does not fall into any of the normal classifications-would be listed as X plus the tracing; if there are more than two deltas, the tracing will be from the extreme left delta to the extreme right delta. Thus, an accidental with an outer trace would be listed as XO.
An ulnar loop would be listed according to its ridge count: 17, for example, if there are 17 ridges between the delta and the core.
A radial loop would be listed according to its ridge count plus 50; thus, a radial loop with 17 ridges between the delta and the core would be listed as 67.
An arch is shown as AA and a tented arch as TT.
Ten Arches and Bunny Feet
Remember that arches and tented arches together make up slightly less than 5 percent of all fingerprints; the proportion actually varies very slightly according to race from about 3.5 percent to about 6.5 percent. Arches and tented arches are most commonly found on the index and middle finger. Very few people have arches on all ten fingers; in fact, in Albany, as our total number of fingerprint cards rose from about 10,000 when I first began working with them to around 13,000 by the time I left, the number of all-arch cards remained constant. There were three. Two of them came from people I never met; the third came from Eddie.
Eddie was probably the stupidest burglar I ever met. Knowing from his first trial—assuming, of course, that he paid any attention to it—that he had the rarest ten-finger classification in the world, he went right on burglarizing, and he went right on not wearing gloves.
And Eddie had bunny feet. When law-enforcement people say somebody has bunny feet, they do not mean it as a compliment. A person who has—or grows—bunny feet is a person who has repeatedly escaped from custody.
Eddie burglarized a grocery store, eight doctors' offices, and several other businesses. We caught him. It wasn't too hard; all those arch patterns helped.
Eddie has another strange habit. When he's caught, he denies everything for about two minu
tes and then he starts to giggle and says, "Well, I guess you got me."
Eddie went to prison.
Eddie grew bunny feet. On his way home, burglarizing in every town he passed through on the way, he stopped and burglarized a sheriffs office in a nearby town.
We caught him. It wasn't too hard; all those arch patterns helped.
Eddie denied everything and then started giggling.
Eddie went to prison again, this time as a habitual offender. A few days later, I met a guard from the prison Eddie had been sent to the second time. I mentioned Eddie and his bunny feet. The guard thanked me for the warning.
I haven't seen Eddie since. But if I ever do, I'll recognize him easily. All those arches help.
Missing or amputated fingers are shown as XX, and patterns so scarred or mutilated that they are unreadable are shown as SR.
Figure 5-5 is a fingerprint card, already classified according to the simple Henry classification, not the FBI extensions. Following is the NCIC classification of that card:
WOTT0305TTDI06TT15TT That formula would allow any fingerprint technician to reconstruct the classification of the fingerprint card. But it would not allow identification.
Identification by Prints
And here's where we get into what fingerprinting is all about.
If fingerprints were composed entirely of straight or curved complete lines, identification would not be possible. But they are not. These lines stop and start, form islands. The main minutiae— identifying points—points of similarity, as identification technicians call them—of fingerprints are ridge endings, dots, short ridges, bifurcations and trifurcations (see Figure 5-6).
It is the arrangement of these identifying points that make identification possible. One of the individual fingerprints I succeeded in memorizing had a long, near-vertical triangle beginning at the delta, which formed the point of one of the angles. Toward the core, it had two long ridges followed by a short ridge. I did not have to memorize the entire print: These few details gave me five