KEN MOSES carried the fingerprint card with him for eight years. The print belonged to a man who had shot and killed the resident of a house he was burglarizing.
To Mr. Moses, a San Francisco Police Department inspector, it was a sad end for a woman who had survived a Nazi concentration camp, immigrated to the United States, and had finally emerged into the middle class.
``It was one of those cases you really wanted to solve,'' Moses says.
From the time he lifted that latent print of the killer in 1978, Moses spent his spare hours trying to match it against thousands of other fingerprint cards in the department's files - to no avail.
Then, in early 1985, San Francisco installed an Automated Fingerprint Identification System (AFIS). The first thing Moses did was to put his pet fingerprint into this computer system. Four minutes and 10 seconds later, the computer had matched the print to another on file. The police arrested a suspect within 24 hours. Faced with what courts consider nearly irrefutable evidence - a positive fingerprint ID - the suspect pleaded guilty to murder. He is now in prison.
Computerization of fingerprinting ``is probably the biggest advancement in law enforcement since the automobile,'' says Tom Wilson, a technology specialist at Search Group Inc., a nonprofit consortium of the 50 states which analyzes criminal justice problems.
Cities and states are putting their fingerprint files on line faster than you can say ``ink pad.'' About 70 law enforcement agencies, including 20 states and the District of Columbia, have systems working or on order. Eventually, law enforcement agencies envision a nationwide, computerized fingerprint network that would leave little escape for those who leave their prints at the scene of a crime. It is a vision that even civil libertarians have few reservations about, since the accuracy of computerized fingerprint matching would reduce the chances that an innocent person would be arrested by mistake.
It will likely take two or three years to iron out the technological wrinkles in automated systems so that different states' computers can communicate with each other, says Mr. Wilson. Within five years, about two-thirds of the population will be covered by AFIS, he says.
Those who are already using the system say AFIS has done more than any other tool to improve law enforcement's ability to catch criminals, send them to jail, of-ten for longer terms, prevent theft and violent crime, and ease the workload on police and prosecutors.
San Francisco, for example, has reduced its burglaries 28 percent since it turned on AFIS in 1984, bringing it below the national average for the first time. In that year, police solved 816 cases involving latent prints (that is, unidentified prints found at the scene of a crime) - a 13-fold increase over the previous year, when 58 such cases were solved.
Houston police cleared more than 600 cases in the first year of AFIS. Prince George's County in Maryland made 150 matches in the computer's first nine months. Police in New Orleans say they have tripled their success rate in matching fingerprints and convicting the criminals.
Most of these cases would not have been pursued without computers, says Gary Cooper, who manages California's $22 million computerized fingerprint system, called CAL-ID. ``Before AFIS, you'd just file it [the latent print] and wait till you got a suspect,'' he says.
The reason, he and others say, is that states and major cities have so many fingerprints on file - California has 8 million - that it would take years, sometimes decades, to try to match a single print manually. With the computer, it takes a few minutes.
Matching these ``latent'' fingerprints translates into a much more dramatic drop in crime, says inspector Moses, since the average burglar commits 100 or more burglaries a year. ``So when you put 1,452 burglars in prison,'' as San Francisco has since 1984, ``you prevent at least 100,000 burglaries the next year.''
Defendants are also more likely to plea-bargain rather than fight the charges, which saves judge and prosecutor time.
``There's very little defense against a fingerprint left at a crime scene,'' notes Wilson at Search Group.
Perhaps the most dramatic use of the system will be in dramatic crime.
In late 1986, New Orleans experienced a sudden wave of violent, random murders. ``We had no leads, and [the killer] didn't seem to have a motive'' for the attacks, says Wayne McKenzie at the New Orleans Police Department. Then, the day after Christmas, the attacker left his print on the car of two of his victims. The fingerprints were entered into the computer, and it identified a suspect within a few minutes. Three days later, the man was in custody.
``He was doing a murder every other day,'' McKenzie says, noting that speed is of the essence where serial crimes are concerned. The man has been convicted of four murders and faces charges on eight others, as well as on 25 armed robberies.
Technology experts say it will take a long time for the biggest fingerprint repository of them all - the Federal Bureau of Investigation - to hook up with the states.
The FBI has 20 million people on file for crimes, and, since most people have 10 fingers, that's 200 million prints. Sheer size is not the problem, however. Rather, the FBI's current system is not organized to conduct a ``cold'' search - that is, searching an unidentified, latent print. It can match a print only when the police have a suspect. The upshot is that the states currently have no centralized way of searching for, say, a serial killer who has crossed state lines.
Dennis Kurre, the head of the FBI's identification division, says the FBI wants to modernize its system so it can do ``cold'' searches for states by 1993. But funding for modernizing the ID division has already been suspended once.
The logical next step is for states to share information about suspects.
But there is a hitch. The three major companies that make automated fingerprint ID systems use different languages and measurements in computerizing the fingerprints. Consequently, Indiana can't exchange information with neighboring Illinois, because they are on different systems.
In December 1986, the National Bureau of Standards approved a standard that would translate data into a format that the various machines could understand. All that was needed was for the vendors to agree on the standard.
``I've been trying to shake loose cooperation for two or three years now,'' says Ray Moore, a scientist at the National Bureau of Standards.
Nick Williams at De La Rue Printrak, one of the vendors, concedes that vendors are reluctant to agree on a standard, ``because it would mean disclosing proprietary information.'' But a little arm-twisting will probably solve the problem: ``The law enforcement community will have to establish the standard and the vendors will follow,'' he says.
And arm-twisting is just what states are beginning to do. This week, 11 states are formally banding together to form the Western Identification Network. Six of the states - Nevada, Montana, Idaho, Wyoming, Oregon, and Utah - are jointly buying a system, which they could not afford individually. Five neighboring states that already have AFIS will be able to ``interface'' with them.
``We're telling the vendors that the system we buy should be able to send information from one to another,'' says Allen D. Jones, chairman of the Western network. To do that, the companies will either have to adopt the NBS standard or develop their own that enables them to interface with other systems.
The move by the Western states will be felt across the country, says Wilson at Search. ``The next time a neighboring state such as Kansas gets a system, they're going to request compatibility with the Western states,'' he says. ``It's going to go down like dominoes.''