I came to think of the left side of the street as the medical section of Tent City. First came two refrigerated trailers, the kind normally pulled along the highway as the back part of a semi or tractor-trailer rig. These trailers, known as “reefers,” were attached to large diesel-powered refrigeration units that periodically roared to life as they tried to keep the trailer's contents cool. Here, workers stored the body bags as they were brought in from the disaster site so that we could process the remains in a steady stream, rather than allowing them to stack up and deteriorate in the heat of the day.

To the right was the “support side” of the street-another group of white tents whose yellow- and blue-striped roofs lent an incongruous circus air to the area. I noticed a chapel, complete with altar, flowers, and rows of benches-a private yet communal place where the rescue workers could share a moment of much-needed quiet. Then there was a tent stacked with mysterious-looking cardboard boxes, which I later found out contained such personal items as toothbrushes, soap, shaving supplies, even packages of underwear and socks-anything that displaced workers might need when they had to work here for days at a stretch, without time for a long commute back to the suburbs.

The next tent was filled with police officers. When I saw some of them standing around drinking coffee and smoking cigarettes, taking a brief break from the urgent pace of work, I felt the urge to smile. I was finally starting to feel at home.

Then there was the morgue, whose open tent flaps allowed a clear view all the way inside. I saw six stainless steel autopsy trays, each long enough to hold a human body, supported by a pair of sawhorses. Behind each tray was a stainless steel cart filled with latex gloves, marking pens, plastic bags of all sizes, scissors, knives, and a bone saw.

A team of medical specialists surrounded each tray, moving in the highly synchronized choreography that you see in an operating room, an ER-or a morgue. The noise was overwhelming as the workers shouted to one another over bone saws that whined like dentists' drills. Clerks wearing hospital scrubs stood by, passing labeled folders to the pathologists each time a new body bag was loaded onto their table. And over it all hung the heavy smell of death, along with diesel fuel from the nearby reefers.

I wanted to stop for a closer look, but my colleagues were making their way to the tent labeled DMORT. Inside, their twenty-by-twenty-foot nylon tent was crammed so full of people and supplies that there was barely room to move. The side walls were stacked with boxes of disposable surgical gloves and gowns, while along the back were paper towels, notebooks, and office supplies. I noticed some military-issue canvas cots where exhausted morgue workers might catch a catnap during breaks, and in the center of the tent were a half-dozen folding chairs surrounding a makeshift table-a large wooden spool normally used to hold a coil of steel cable. It remained me of something you'd have at a fishing camp.

This table held the sign-in sheet that each of us had to initial each day before we scattered to our various assignments. The DMORT supervisor on duty, Cliff Oldfield, knew that this was my first day, so he told me to stand by. In just a few minutes, he promised, he'd tell me everything I needed to know.

When Cliff brought me over to the morgue, Amy Zelson Mundorff, the forensic anthropologist who worked full-time for the city's chief medical examiner, was ready to greet me. I saw immediately how personal this tragedy was for her and her colleagues-although she was smiling brightly, large purple and green bruises ringed her eyes, and her forehead sprouted a lump large enough to cast a shadow. Cliff had told me that she, along with Dr. Hirsch and several other OCME staff members, had rushed to the Twin Towers shortly after the first plane hit. They were in the process of trying to establish a site for a temporary morgue when Tower 2 had come crashing down. The blast blew her headfirst into the marble pillar of a nearby building, while Dr. Hirsch escaped death by inches. One staff member suffered a massive concussion and fractured ribs, and another's leg was shattered, leaving bone and muscle exposed to the air.

However, Amy, just barely five feet tall and sporting a head full of curly black hair, was still very much alive and eager to get on with the business of victim identification. After she gave me a whirlwind tour of the medical examiner's office, we wound up at ground level in the morgue tent, where she took her place at the first autopsy table, the one designated for triage. As she worked, she quickly explained the overall recovery and identification protocol.

The protocol we were using was based on the emergency plans that the OCME had developed well in advance of September 11, with modifications that Dr. Hirsch had added when the scope of the disaster became clear. This protocol fascinated me: It was both different from and similar to the setup at the other incidents I had worked.

Most mass-fatality morgues are set up in sort of an assembly-line format. Rescue personnel bring in bodies or body parts to be photographed, x-rayed, and preliminarily identified by careful scrutiny of superficial characteristics such as hair color, skin color, clothing, and perhaps jewelry.

If the body is relatively intact, the forensic pathologist will conduct an exam that is not too different from normal autopsy protocol. He or she will photograph, weigh, measure, and describe the body in painstaking detail, including observations on its overall condition and any old scars or other evidence of surgery that the preliminary exam might have missed. The pathologist will also document the acute injuries that most likely caused the person's death-documentation that would eventually include photographs and a detailed written or dictated description.

In some cases, as in the Oklahoma City bombing, living perpetrators are involved who may eventually stand trial for murder. The medicolegal details of autopsies might well have ramifications for their criminal prosecution.

But the attack on the Twin Towers was a completely different situation. There was no reason to conduct traditional medicolegal autopsies: Millions of people had witnessed the events as they happened. We all knew that these were deaths by homicide and that none of the hijackers could have escaped alive. It would be virtually impossible to identify the specific causes of death for each victim. All that mattered was identifying the victims themselves-but that task was difficult enough.

Here, the first step in the identification process was triage, a French word that means “to separate.” The term is used primarily in battlefield or medical situations, where patients or victims are separated into categories based on how urgent it is to treat them. This triage process began with a forensic anthropologist who had to identify and separate every single bit of human tissue that came through the morgue door. If a bag came in filled with a twenty-pound mass of muscle, skin, and bone, we had to be able to tell either by feel or by sight if the tissues were connected. If a bone led to a tendon, and then to a muscle, and the other end of the same muscle was connected to yet another bone, then the entire specimen could stay together and be processed as a single set of remains, because the “connections” established the fact that these were the remains of one person. But if we determined that there was no physical connection between one part and another, then we'd separate the remains and give them two different case numbers. Hopefully, DNA analysis could later link the two, but for now we kept them separate.

Triage could also reunite body parts. If we found two ends or sides of a fractured bone in this mass of tissue, the broken surfaces of which we could match or “reapproximate,” then we could say with all confidence that these two pieces were from the same victim. Our ultimate goal was to match all of the remains with a name, but much of that would come later, through DNA analysis.