In the first gray light of dawn, a police criminologist crawls along a road on his hands and knees, peering for paint chips or shards of glass from a mystery vehicle that fatally struck a pedestrian the night before.

In a state-of-the-art laboratory, a mechanical engineer painstakingly attempts to recreate a car fire that claimed the life of a motorist.

At a computer keyboard, a reconstructionist uses sophisticated software and digitally scanned photographs to map a crash scene in three dimensions and deduce who hit whom, how and when.

Scenes from the hit television show CSI, perhaps? Not at all. The incidents above involve real-life professionals working on cases with real-world consequences for victims, their families and the public. On TV, CSI has risen to the top of the ratings by depicting smart people doing cool, high-tech stuff to solve crimes. In the real world, crash scene investigators and accident reconstructionists also do cool, high-tech stuff. But unlike their fictionalized counterparts on the tube, they labor in relative obscurity. And their work has very real, often life-or-death consequences.

What exactly do accident investigators do? How accurately does TV portray them and their work? And how do they make roads safer for everyone? As it turns out, the answers give you a glimpse of real-world professionals that inspire the television dramas and provide some valuable lessons that can make you a better driver.

Piecing Together Puzzles
TV focuses on only one particular portion of the profession--investigators working for law enforcement on possible criminal cases. In fact, the accident investigation field also includes professionals with similar training and expertise working for private firms. They take cases for insurance companies, parties in civil suits and other clients. And unlike the police, these professionals may begin their investigations days or even weeks after a crash occurs.

Whether they work for police departments or private firms, accident investigators have two basic job titles, depending on the nature and level of their training.

"Crash scene investigators" photograph the site and the final resting place of the vehicles involved, take elaborate measurements and collect evidence. They examine and photograph tire marks, document gouges on trees and curbs and pavement, collect blood samples from the vehicle and elsewhere, and otherwise gather all the pieces of the puzzle.

Crash scene investigators then turn over all the pieces to "accident reconstructionists." Although they might sometimes also visit the scene to gather evidence, accident reconstructionists have special training in math, physics, mechanical engineering and other disciplines.

"An accident reconstructionist does all the analytical work," explains Haseeb Ghumman, managing engineer with Accident Reconstruction Plus, a private firm headquartered near Buffalo, N.Y. "Basically, we put all the pieces of the puzzle together." The completed puzzle gives a picture of what actually happened--the sequence of events, the vehicles' paths of travel, their speeds at various points in time and ultimate results.

Tales Tires Tell
One of the biggest challenges facing accident investigators is sorting relevant evidence from irrelevant evidence. Take tire marks, for instance. To a trained analyst, tire marks can reveal points at which a vehicle braked, accelerated or changed direction. On a curve, tires leave characteristic impressions -- "yaw marks" -- that can even offer clues to the vehicle's speed at that point. And you might find them 100 feet or more from the eventual point of impact.

"Tire marks don't live forever," explains Grahme Fischer, president of Technical Problem Solvers, Inc., and director of the New York State Traffic Accident Reconstructionists Society (NYSTARS).

"Most disappear within a month. In the city of New York, you might see 10 spurious marks for each collision." The faster investigators get to the scene--and the more isolated the location--the better.

The same goes for other types of evidence--paint chips, grass fragments, automotive fluid residue and vehicle parts thrown off in the force of a collision.

"You find all kinds of junk on the road," says Bob Genna, director of the Suffolk County Crime Lab. He uses location and other clues to spot the potentially relevant bits. "You tend to find particles resulting from a recent collision toward the travel portion of the roadway, before they're thrown over by passing traffic or swept off by the wind," he says. "And the fracture itself can tell you whether it's fresh. The surface isn't worn and doesn't have material adhering to it."

Who Was Driving?
Investigators pore over other evidence as well. For example, if a crash involves fatalities, the victims' clothes may offer clues." Clothing will often pick up plastic content from the seats or dashboard or other interior surfaces," explains Fischer. "And seat belt material may transfer as the belt is stretched." Characteristic stretch marks on a safety belt, impressions left on an occupant's body and even blood splatters can help determine people's positions in the vehicle.

Positions become critically important in cases where some occupants survive and others don't. A survivor may deny being the driver to escape criminal penalties or legal liability. And with no other survivors from the vehicle, who can contradict the story? The evidence can--and often does.

Accident reconstructionists use sophisticated tools to assemble and analyze all this data. At Accident Reconstruction Plus, for instance, Ghumman maps sites with a laser-scanning instrument used by surveyors. It scans a site from three or four different angles, and a computer later assembles the scans into a 3-D map. "It's neat, fast, accurate and remote," says Ghumman. "You can set it up on a tripod at the side of the road, so you don't have to disrupt traffic."

Event data recorders--so-called "black boxes" on vehicles that record their speed, the driver's use of controls and other data in the seconds leading up to a crash-add to the information available. Ultimately, though, all the technology in the world can't replace old-fashioned eyes, hands and brains."You can't replace the human factor," says Ghumman. "If the data is collected correctly, then the analysis will yield correct results."

Not "As Seen on TV"
Above all, a detailed accident reconstruction takes time. Most accident investigators cite this element as the biggest difference between their real-life work and the TV portrayals."

Television gives people the perception that things get done very quickly," says Genna. "In fact it may take several weeks to develop photos, map the scene, analyze the evidence, make consultations and write a report."

Real-life professionals raise a few other quibbles with TV shows as well. For one, they notice unrealistically young actors portraying investigators and reconstructionists. "This isn't simple work," says Fischer. "It requires a mature engineer with 15 or 20 years of experience in diversified fields. From what I've seen, the actors on TV look far too young."

Genna also has noticed one or two procedures used on TV that would never be used on the street--for instance, taking the liver temperature of a body at a crime scene to determine time of death. "We would take a rectal temp and never damage the body," he says. Otherwise, Genna finds the TV shows fairly accurate. "I think they're great," he says, "because they introduce the general public to a field no one knew existed 10 or 15 years ago, at least pre-OJ [Simpson]."

Lessons Learned
As you might expect, looking at vehicular carnage day after day affects these professionals' behind-the-wheel behavior. Every single investigator and analyst interviewed for this story claims to drive slower than average. As Genna puts it, "I'm not saying that I stay in the granny lane all the time, but I get passed a lot. I tend to drive very, very slowly down residential streets in particular, and I'm especially cognizant of pedestrians."

"Everyone who has worked in this office for two months or more wears a seat belt, whether or not they wore one before they started," says Fischer. "Personally, I also avoid driving near or directly in front of big trucks. I avoid them with the knowledge that a significant fraction of trucks have maladjusted brakes, and they're not capable of stopping nearly as well as they should. I've seen the mayhem that can result. Let's just say I don't stay in front of big trucks for long, and when I pass them, I do it in a hurry."

Fischer also carries a camera in his glove compartment, and he recommends every driver do the same. "One of those cheap, throw-away cameras is perfectly adequate," he says. If you're ever involved in a collision--no matter how minor--he recommends taking photos of the vehicle on all four sides, close-ups of the damage (or lack of damage), and features of the roadway.

"There is a small percentage of people who are in the business of creating the impression of injury," he says. "Or they simply believe that because they've had an accident, someone should pay. What you think is an inconsequential collision may turn into a lawsuit six months later. The apparent lack of severity at the scene is no indication of how serious it may become. And photos provide the best objective evidence."

Hunting Down the Truth
Despite the distinct undercurrent of cynicism in his advice, Fischer remains firmly convinced that his profession contributes to traffic safety. "I certainly believe that economic pressure from lawsuits provides an incentive to manufacturers to design safer cars," he says. "Accident reconstructionists make a big difference in product liability cases in which manufacturers did not implement appropriate safety measures."

As for criminal investigations, Genna takes special satisfaction in investigating hit-skip collisions involving pedestrians.

Genna recalls one hit-and-run incident a few years back, when he got a call in the wee hours of the morning. He hung around the scene for hours, until sunrise, when he could search for evidence in better light. Just after dawn, he spied a few tiny paint flecks on the road. The paint--combined with the pattern of injuries on the victim--suggested a white SUV. Based on his evidence, police issued a "be on the lookout" that eventually led to finding the vehicle in the perpetrator's driveway, with tell-tale damage to its front end.

"It's satisfying when you find something that starts a snowball effect like that," he says. "I'm especially proud of the work we did on that case."

Once, Technical Problem Solvers' Fischer heard a prominent anthropologist speak about the perfect job for which humans are adapted. "He said that you would go out hunting, searching for animal tracks," explains Fischer. "You'd find them, follow them and come upon an animal. You'd kill it, then bring it back to the village, and everyone could eat."

"So the perfect job involves working independently, using your intelligence, relying on your senses and using your physical skills," continues Fischer. "Basically, this anthropologist was describing detective work. I think I have that job."

Few people who've watched reconstructionists in action--either on TV or in real life--would disagree.