CHAPTER 6 FINGERPRINT ANALYSIS

• Jurisdiction: United States

Chapter 6 Fingerprint Analysis

Overview

The science of fingerprint identification rests upon two hypotheses:

• All fingerprints are permanent and unique and therefore can identify one individual compared with another.
• By visually examining certain detail on less than a complete fingerprint or a completely clear fingerprint, examiners can conclusively identify a subject.

Fingerprints are formed in utero by the fingers of a fetus moving through amniotic fluid. For this reason, even the prints of identical twins, who share the same DNA, are different. Most scientists agree that fingerprints are unique, that they do not change over time, and that they cannot be easily altered. A typical fingerprint has approximately 200 distinguishing features, called ridge characteristics, which can appear in any one of a number of configurations. Fingerprint identification uses a combination of automated search based on an algorithm coupled with visual inspection and confirmation, called the ACE-V approach (Analysis, Comparison, Evaluation, and Verification). The basis for matching requires a conclusion that a latent fingerprint lifted from a crime scene have no visible characteristics that differ from an exemplar print taken from a suspect and possess a number of characteristics that appear identical in size, shape, and location.

Fingerprint identification has been accepted as forensic evidence since the early 1900s.¹ Not until quite recently has there been any challenge to its reliability. However, a few scholars as well as one federal district court argued that fingerprint identification did not pass the Daubert tests for reliability because no scientific studies had been conducted outside the fingerprint community to verify the required minimum number of comparison points to declare a match.²

This 2002 case was later retracted, but it spawned a number of subsequent challenges to fingerprint identification under Daubert and similar state statutes. These challenges uniformly failed.

The NRC Report in 2009, however, led to another spate of challenges, as it concluded that fingerprint identification needed better documentation for each step of the ACE-V process and that error rate needed further study:

[A]dditional research is also needed into ridge flow and crease pattern distributions on the hands and feet. This information could be used to limit the possible donor population of a particular print in a statistical approach ... Additionally, more research is needed regarding the discriminating value of the various ridge formations and clusters of ridge formations.³
The NRC Report also called for more research on the various factors that affect the quality of latent prints such as condition of the skin, residue, and mechanics of touch. And because the print examiner uses his or her subjective judgment in declaring a match between a latent print and an exemplar from a suspect, the NRC also recommended that the examiner document the basis for subjective conclusions.

Although challenges based on the NRC Report that fingerprint matching fails Daubert have not succeeded, they have led at least one commentator to criticize the NRC Report and refute its contentions about problems with fingerprinting.⁴

There have been almost no reported cases of fingerprint identification errors, other than mechanical errors such as copying the wrong person's name on a file. One highly publicized fingerprint identification error took place in 2004, in which a fingerprint that appeared on a piece of luggage in Madrid at the site of a train station bombing was incorrectly reported as identifying an attorney in Seattle, Washington, who happened to be a Muslim.⁵ Most people believe this error was the result more of cutting corners in the haste to find a terrorist than any fault in fingerprint identification methods.

Chapter Objectives

Based on this chapter, students will be able to:

1. Explain the scientific principle underlying fingerprint identification.
2. Appreciate the difference between class and individual characteristics of fingerprints and the three levels of detail in fingerprints.
3. Understand the process of lifting and comparing latent and exemplar fingerprints.
4. Explain the ACE-V process.
5. Understand the use of computers using an AFIS database.
6. Explain the history of using fingerprints in court.
7. Apply each of the Daubert tests to fingerprint identification.
8. Evaluate the NRC recommendations for further fingerprint research and form an opinion as to whether these recommendations are valid.
9. Identify errors in the initial identification of fingerprints in the Madrid bombing case.

The Evolution of Fingerprint Identification

Fingerprinting is one of the oldest forms of forensic evidence, and has been routinely admitted in court to prove identity. Its predecessor was anthropometry, which is discussed in Chapter 4, "The Scientific Method and 'Junk Science.'" Anthropometry was a system developed by Alphonse Bertillon in which measurements of body dimensions, including the size of ears, were thought to create a unique profile of an individual's physical characteristics that could identify him to the exclusion of all others. Although anthropometry did not prove discriminating enough for this task, the use of fingerprint comparison is based on the same principle: that no two people share the exact same fingerprints.

Scientists today agree that fingerprints are unique. This is based, in part, on the knowledge that fingerprints are formed in utero by the movement of the fetus' hands through amniotic fluid. For this reason, the fingerprints of even identical twins (who share the same DNA) are different, which shows that fingerprints are not a genetic trait. The uniqueness of fingerprints was confirmed by a government study in which 50,000 fingerprints consisting of all loop patterns from all white males were compared. The study found that no two were identical:

The goal of this study, which was comprised of two separate tests, was to determine the probability that fingerprints of two people could be identical. Donald Ziesig, an algorithmist at Lockheed Martin Information Systems who played an important role in developing the FBI's computer-based fingerprint system (the Automated Fingerprint Identification System, or AFIS), was a developer of the 50k x 50k study and explained in detail how it operated. The result of the first test, in which full-sized, one inch fingerprints were compared with each other, was that the probability of finding two people with identical fingerprints was one in ten to the ninety-seventh power.
In the second test, the rolled prints were artificially cropped to the average size of latent prints so that only the center 21.7% of the rolled prints was analyzed, with the resultant conclusion that the probability of finding two different, partial fingerprints to be identical was one in ten to the twenty-seventh power.⁶

Loops

Characterized by one or more free recurving friction ridges and one delta. (When the hand from which the loop pattern originated is known, you may determine if the recurving ridges originate from the little finger side (ulnar loop) or the thumb side (radial loop).)

Arches

Characterized by friction ridges lying one above the other in a general arching formation.

Whorls

Characterized by one or more free recurving friction ridges and two points of delta.

Ridge Characteristics

1. Bifurcation — The point at which one friction ridge divides into two friction ridges

2. Enclosure — A single friction ridge that bifurcates and rejoins after a short course and continues as a single friction ridge

3. Ending Ridge — A single friction ridge that terminates within the friction ridge structure

4. Short Ridge — A single friction ridge that only travels a short distance before terminating

5. Ridge Dot — An isolated ridge unit whose length approximates its width in size © 2004 WARD'S Natural Science

Some critics noted that this sample was restricted, and did not cover a wide spectrum of the population. However, no one has disproved the hypothesis that fingerprints are unique. Scientists have also concluded that most wounds will not eradicate a fingerprint; at most, a deep cut into the dermis layer may result in a scar, but the rest of the fingerprint will remain the same.

If you look at your own fingers, you will be able to see that each finger bears one of three primary patterns: loop, arch, or whorl, of which there are subgroups. There are two types of arches and loops and four types of whorls. There have been a number of studies done to determine the distribution of these major patterns in the population. In general, loop pattern is most common — about 60% of people have this pattern. The whorl accounts for about 35% and the arch for about 5%.⁷ These characteristics are therefore "class" characteristics. You can be excluded if your fingers all have loops and the print from the crime scene is a whorl. But if the crime scene print is a loop, that alone does not mean it is yours. It could belong is 60 out of 100 people chosen at random. Fingerprints also have fixed reference points called "deltas," seen in loops and whorls, and "cores" seen in loops.

The principle that can identify a person in fingerprint analysis depends upon much smaller characteristics that are called ridge characteristics or Galton points (after Sir Francis Galton, who identified them in the late 1800s). These are the patterns made by the raised lines on each print. Each fingerprint has from between 75 and 175 of these ridge characteristics on each finger.⁸ Fingerprint examiners refer to three primary types of ridge characteristics: ending ridge, bifurcation, and ridge dot. Variations, such as an enclosure, island, crossover, bridge, or trifurcation are combinations of the three major patterns. The chart below taken from Ward's Natural Sciences includes five ridge characteristics:

• Bifurcation—where one friction ridge divides into two.
•

...