In June 2003, the Washington, D.C., metropolitan region-the states of Maryland and Virginia, and Washington, D.C.-was about to experience terror again. These communities dealt with the events of September 11, 2001; the D.C. sniper in October 2002; and now they would have to deal with a serial arsonist. This arson-related reign of terror lasted almost two years and affected the lives of everyone in the region.


The possibility that an arsonist might be operating in the region came to light through a simple promotional exam. The commander in the Prince George’s County (MD) Fire Investigation Unit was comparing notes with a commander in the Washington (D.C.) Fire Investigation Unit during a promotional exam. They noticed that they had been experiencing a series of similar-type fires. This is not that unusual for these two departments, since they border each other and respond to thousands of working fires each year. Both deal with arson on a large scale.

However, the difference in this case was that these fires seemed to be almost, if not exactly, the same:

  • They occurred in the early morning hours between 2 a.m. and 6 a.m.
  • They were set at single-family homes.
  • They were set at or near an entrance area on a porch or deck.
  • They were in similar types of neighborhoods.

Both investigators agreed to return to their offices and study the issues more closely and to meet again in a few days. In the meantime, Prince George’s County (PGC) experienced a couple more of those fires. Working proactively, investigators contacted agents from the Bureau of Alcohol, Tobacco and Firearms (ATF) and asked them to come out and look at the information they had. In addition, they took all the evidence they had collected to the National ATF lab in Ammendale, Maryland.

While they were waiting for word from the lab, Washington, D.C. experienced another fire and requested ATF assistance. While investigators were reviewing the evidence at the last D.C. fire, the lab called and confirmed that the materials brought in from the PGC fires were the same. The evidence from the most recent D.C. fire was taken to the lab and compared with the material from PGC. Again, the lab found the materials to be similar.


The lab confirmed the following:

  • Some type of polyethylene (plastic) jug was present.
  • The container was consistent with a gallon container.
  • Some type of wick material was present.
  • Some type of plastic bag was present.
  • A flammable liquid (gasoline) was present.


Investigators quickly brought together administrators from the D.C. and county police and fire departments and formed a task force. The task force agreed that it would proceed as follows:

  • Full-time investigators would be assigned.
  • PGC Fire would provide space.
  • A tip line would be established.
  • The ATF National Lab would be the single lab source.
  • A behavioral and geographic profiler would be consulted.
  • All reports would be prepared in an ATF database.
  • The U.S. Marshal Service would deputize all task force investigators.
  • A reward fund would be established.
  • County Fire Chief Ron Blackwell would be the voice of the task force and would handle all media requests.
  • The executive (senior) staff would make all decisions.
  • Regular meetings would be held with senior staff and attorneys from D.C. and Maryland.
  • Task force commanders would prepare weekly briefing reports for the ATF senior staff.

Once the task force was set up, commanders knew additional people would be needed to assist in the investigation. The following agencies participated initially: the D.C. Police, PGC Police, Maryland State Fire Marshals, Maryland State Police, PGC Sheriffs Office, and the Anne Arundel County (MD) Fire Department.

The increase in staff necessitated more space. The county fire unit made a training room in its office available for investigators. As a result, command staff and investigators had to work in separate areas to allow the county investigators to keep their workspace active. The separation caused hard feelings among some investigators. Anytime they needed to gain entry to the space where the command staff was, they had to knock to gain entry. This created an atmosphere of “us vs. them,”-the grunts vs. the bosses. The problem continued until the task force changed its method of operation: It established a unified command and divided the task force into sections, each with a supervisor. This system enabled supervisory staff and investigators to work in the same workspace and afforded them the interaction they needed.


Operational guidelines were established and disseminated. The first and possibly most important task was communication. The events of September 11, 2001, taught us that reliable communication among multiple agencies was of paramount importance. Since we had no single system that would allow us to do this, we provided all task force members who did not already have one with a cell phone. All investigators’ information was entered into a single database so that everyone could receive text messages by phone. As a means of backup, pager messages were sent simultaneously. The phones also allowed direct radio contact for all personnel.

Even though the cell phones and pagers were in place, messages were not always noticed once they were sent. A process was needed to disseminate information to all task force members. The commanders from jurisdictions where fires occurred contacted the other commanders in the task force. They, in turn, contacted all the members in their group. It was critical that all members’ contact information be kept current. The command center, staffed 24 hours a day, sent a page as a backup. One person was on duty at all times when the command center operations were shut down, so tip lines could be answered, the phone could be answered, and pages could be sent in and research done for field investigators.

We are most grateful to those who worked in the command post overnight. The job they performed may have gone unnoticed in their view, but those in the field appreciated all they did and felt good knowing that they could be called on when they needed something.


A two-tier process was established for working fire scenes. The first and most important was our evidence group. The same people had to be involved in the collection of evidence every time there was a fire. The evidence team met with investigators from the agency where the fire occurred. They discussed what they had, went to the fire area together and dug it. (“Dug” refers to the process in which fire investigators go through each layer of burned debris looking for the origin and cause of the fire. These layers must be dug by hand and looked at very carefully. Investigators are also able to find more detailed pieces of evidence during this process.)

The host jurisdiction collected any evidence found, bagged and tagged it, and then turned it over to the task force. This enabled the chain of custody to remain in place, which was important in case the fire turned out to be unrelated to ours. It also kept each jurisdiction directly involved in the investigation and let them know that the task force was not trying to take over but was working with them as a team.

The evidence team also decided that it would use an accelerant K-9 on every scene to help determine the presence of a flammable liquid. Three handlers were assigned to the task force; when one was not available, the county had committed to making one of its handlers available. The process worked like a charm for the most part. These problems were quickly addressed, keeping the operation intact.


The investigation group, the second tier of our scene investigation process, conducted interviews with the victims, did the neighborhood canvass work, visited the business locations in the area, determined who had video security and obtained all the tapes, and determined which gas stations had sold small amounts of product in our target time frame. The group developed a questionnaire for the victims; they developed a canvass pattern that involved doing a two- to four-block crisscross. It worked well. Doing this and putting it on paper made it easy when new people came to the group. They could simply follow the script.


June 30, 2003. Less than a full week from the formal formation of the task force, we had our first callout. It was 0357 hours, and a fire had taken place at 2505 Randolph Street in D.C. The fire was set on the front porch near the door. The process we developed for scene work went well. All the pieces of our device were present. Our arsonist was hard at work.

Prior to the task force’s having been set up, 24 fires had been set-interestingly, 12 in D.C. and 12 in PGC. Each of these scenes would have to be reviewed a second time, since in some cases no evidence was collected on the initial investigation. This was not a blemish on the first investigators; they did not know a device was present and, because of the nature of the fires, they could easily have thought the device was just part of melted siding on the homes. This was a major challenge for our evidence team; it would have to visit all of the original scenes and determine if any evidence was present and get the investigators from the host jurisdiction to come back out and package everything according to our agreed-on process. Although it was a major challenge, they handled it well and got all the scenes done in a timely fashion.


We needed to know more about our device. From all of the people we had talked with and from all the research we had done on the Internet, we could not find any information involving the placement of a plastic device. Sure, everyone has seen the standard Molotov cocktail, but that is thrown at the target. We had all seen a scene where a jug was used to pour from, but we had nothing on the placement of a plastic device. The ATF lab was consulted.

With assistance from PGC Fire and its investigations section, we conducted some test burns on a home that was going to be torn down. These burns allowed the unit to see how the device reacted and the kind of damage it created. Next, we conducted about 20 burns in the ATF lab. One-gallon plastic jugs were filled to various levels with gasoline, a cloth material wick was stuck into the neck, and the devices were burned. All of the burns were videotaped. What we found was interesting.

The devices all burned with a slow, low flame just like a candle. It took anywhere from 12 to 20 minutes for the jug to melt enough for there to be free burning, and once the free-burning event took place, the fire would burn itself out quickly. We also found that the bottoms of the jug remained and determined that while the gasoline was burning, it also kept the bottom cool and allowed it to survive.


We went for the entire months of July and August 2003 without a fire. Where was our arsonist? Who was our arsonist? On vacation? Arrested? Killed? The task force spent these two months trying to answer these questions, and more. It was also the first time we had to address administrators’ questions about taking staff away: There were no fires; surely we could send some people back.

Administrators should think about this question now and consider it carefully. When you are sending someone to a task force or establishing your own task force, will you have the commitment needed? You must allow whomever you assign to remain with the unit until all the work has been completed. Here are other questions to ask yourself: Whom would you send to serve on a task force-your best investigator or your dead wood? Is it better to allow the task force to deal with your problem child or children to get them out of your hair for a while, or would it be better to send your best people who will represent your agency in the best manner and provide the task force with useful investigators? A piece of dead wood is dead no matter where it sits.


Near the end of July, Wshington, D.C. began to have a series of fires in the Marina View Towers high-rise residential building. The fires had some of the characteristics of our fires. We were not sure if this was our person who was just making some changes or if we had a copycat. The building was placed on 24-hour surveillance. We developed a suspect and began to watch his movements. He was shopping at the stores in which we had an interest; he was purchasing many one-gallon jugs. We were hopeful this was the guy.

July 29, 2003. There was another fire in the building, and we had investigators on-site. They visited the apartment of our suspect, wishing to interview him. He answered his door. Investigators saw fresh burns on his hands. Search and arrest warrants were obtained. Many plastic jugs were found on the same floor as his apartment. Many news articles about our arsonist were found in the apartment.

We arrested a 57-year-old black male by the name of Paul DuBois. He was charged in D.C. Superior Court with the following: assault with the intent to kill, second-degree burglary, and arson and obstruction of justice. A trial was scheduled for April 16, 2004.

Everyone was happy because the arsonist was caught. During the course of our investigation of DuBois, investigators uncovered information that may have involved him in a double homicide and turned the information over to New York State Police. Team members were getting ready to return to their respective agencies. It had been 64 days, and there were no additional fires. Did we have the guy?


September 4, 2003. A fire was set in a home at 5101 Barnaby Run Drive in Oxon Hill, Prince George County, MD, at 0315 hours. The device was placed at the front of the house near the front door. The evidence team found all the components of our device. We knew our arsonist was not Paul DuBois. Welcome home! Sixty-four days since the last fire. Now just a few days before school starts up again from summer break, we had a fire. Interesting! Some may even call this a clue.

We now decided we must be ready to catch this person and assign the entire task force to night operations. Staff was set up in teams and asked to ride in the city and PGC. The D.C. Police Department assigned a team of 12 officers from different sections to work only fire scenes in D.C. These officers worked the midnight shift and roamed all of D.C. Although the numbers assigned to this group dropped over the course of the investigation, they were a valued and reliable resource.

September 8, 2003. The first day of school and just four days since our last fire, the arsonist struck again. The address was 202 Quackenbos Street in D.C.; the fire call came in at 0349 hours. It was on the front porch near the door, and all the pieces of the device were found. The overnight crews were riding primarily in the northeast section of D.C. This fire took place in the northwest section. Was someone telling this guy what we were doing?

September 10, 2003. At 0237 hours, another fire was set, at 4713 Dix Street in Southeast D.C. The fire was on the front porch by the door. All the components were present. Investigators were out in the area but came up with zero.

The media began to really turn up the heat. Not only were the local papers writing about the case, but every TV news station was running stories. It was on the radio and even on the Internet. Who is the arsonist? Why isn’t anyone seeing him? How is he getting away? Why aren’t people giving investigators information?

September 14, 2003. At 0929 hours, a concerned mother called the D.C. police. She reported that her sons had had an encounter the night before with a man who left behind some type of jug with gasoline in it. The task force was immediately called, and the team was ordered to the scene. The location was 4115 Anacostia Avenue, NE. When investigators arrived, they were presented with the following story:

Three brothers were coming home from a nightclub around 0400 hours. As they pulled up to the front of their house, they saw a man sitting on their front porch. When he saw them, he walked down the sidewalk and approached the car. The passenger rolled down his window. The stranger said he was trying to locate a friend and provided a name. The boys said they didn’t know anyone by that name, and the man walked off.

The boys decided they needed to get a better handle on what this guy was doing and decided to square the block and see what he was going to do. They squared the block, and when they hit their street again, the man was gone. They parked the car and approached the house. When they got to the front porch, they found two plastic bags that contained a one-gallon plastic jug filled with some type of liquid. They also saw a piece of black cloth material. They smelled the liquid; it smelled like gasoline. They quickly entered their home to make sure the home and everyone in it was okay. Their check was completed, and all was okay.

They got back to the porch and decided they needed to get rid of the stuff. They crossed the street to a school playground and park and put it in a trashcan. They then decided that is was not the best place. They took it to a sewer, dumped the gas, and threw the rest in behind it. They stayed awake, keeping vigil in case this guy returned. In the morning, they told their mother what had happened.

The boys were quickly separated and removed to the command post for further questioning. We contacted the Maryland State Police; an E-FITS operator was sent to assist in making a sketch. E-FITS is a computer-based program that assists in making sketch-based photos. The task force now had a sketch of a person of interest. Investigators at the scene worked more than 14 hours. They collected everything, including the iron handrails on the front porch. They opened the sewer and retrieved the jug, the bags, and the cloth material. The ATF lab technicians were picked up at their homes and brought to the scene to handle everything. Everyone was ready for a good night’s sleep.

The month continued, and there were no additional fires. Geographic and behavioral profilers were brought in, and they provided the following information:

  • He may abuse alcohol.
  • He may engage in reckless driving or unsafe practices at work.
  • He may have a criminal history.
  • When things go wrong in his life, he likely projects blame on others. It is never his fault.
  • He likely lies a great deal, and his dishonesty causes him problems in interpersonal relationships.
  • He lacks what would be considered normal empathy or concern for the welfare of others.
  • He may work in a minimal wage position.
  • He may live close to where the fires are being set.


The above information, along with a scripted plea to the arsonist, was made public at a press conference held at the PGC Fire Services Administration Building. Chief Blackwell revealed the sketch of “someone of interest.” He announced a reward of $25,000 and read the plea. We talked many times during the case of doing this type of plea again, but we never did.

The plea is a very powerful tool and should be seriously considered in cases like this. When we finally arrested our suspect, he told us during interviews that the plea was one of the things he thought about the most during the case. He said that while the plea was being made, he even picked up the phone and wanted to call several times. One of the first requests he made after his arrest was to meet with Chief Blackwell. Although the public plea is a great tool, I strongly suggest you consult with the professionals on this issue before you use it, and be sure to consult with them if you decide to use it more than once. There is a chance that the plea can upset and even anger the individual, leading to the setting of another fire or a repeat of the offense for which he is being sought. All this must be evaluated beforehand.

October 8, 2003. At 0414 hours, there was a fire at 1315 Otis Street, in the Northeast (NE) section of Washington, D.C. The fire was on the front porch at the front door, and investigators from the task force were in the block before the fire apparatus arrived. All the components of the device were present. The blocks around the scene were sealed off; everyone out was stopped and questioned. Zero. The guy vanished again.

The fires stopped, but the media coverage picked up. Fire and police departments are in the news every day. We went from being in the news to being the news. Every agency was putting its own spin on the case. Every place you went people were talking about the case. Everyone knew something or at least thought they did. Everyone was talking. Each news agency seemed to have some piece of the case, but they were all different. Investigators were beginning to talk. What in the world was going on? Everyone was waiting for the shoe to drop. It finally happened. The news began with a great carrot dangling, telling viewers to stay tuned and they would be told how the arsonist was setting the fires. WHAT?


The story began with the reporter’s standing in front of a burned-out home. He was standing on the front porch and telling the viewers, “This is how he does it.” What the reporter did next surprised everyone on the task force. He held up a one-gallon plastic milk jug filled with a yellow-looking fluid. He then told the viewers that the arsonist takes some type of cloth material, like a sock, makes a wick, lights the wick (he was actually holding a lighter in his hand), places the device, and walks away.

The following morning at the task force briefing, things were stressful to say the least. Some wanted to arrest the news reporter. He had in fact made an incendiary device. Right? Was the substance really gasoline? Some wanted to summon him to appear before a grand jury and have him tell us where he got the information. Others wanted to remove everyone and bring in new staff. It was simply a mess. Fingers were pointing in every direction. Trust me, no person or agency was left out. In the end, no one person or agency was found to have provided any information to anyone. There had been several personnel transfers, and many, right or wrong, thought they were the results of an information leak. It’s something we may never know. Leaks are part of a long-term investigation, but you cannot let them direct your investigation. It had been 34 days since our last fire, and we did not know if this news broadcast was causing the perpetrator to lie low. Did he think we knew too much? The calls on the tip line did not slow down. Investigators began to develop their own ideas.


We needed a positive lift and got it. The lab results came back from the Anacostia incident. A hair was recovered from one of the bags. We learned that our subject was a black male, perhaps of mixed race. We had DNA. Everyone got his or her focus back.

November 11, 2003. There was another fire. This one took place at 0230 hours at 1700 24th Street, NE, D.C. The device was placed in front of the home near the left-hand corner. All of our components were present. Investigators were again quickly on the scene; they arrived as the fire apparatus arrived. Still, they came up with nothing.

November 16, 2003. The case took a major turn. The arsonist had set 34 fires that we knew about. Twenty-one of those fires were set in the city, and 13 in the Prince Georges County. At 0451 hours on a Sunday morning, a fire took place in Alexandria, Virginia, at 4410 West Braddock Road. The home was a single-family residence on the grounds of a nursing home, where out-of-town nurses lived. In every criminal investigation, there are turns; in arson investigations, the turns generally have a more in-depth meaning. The arsonist usually becomes bored and needs a change to keep up the interest. Was this it? We now had fires in Maryland, D.C., and Virginia, just like in the sniper case.

December 20, 2003. At 0437 hours, a fire occurred in the New Carrollton section of PGC, at 5702 83rd Avenue. The device was placed at the front of the house by the garage door. What was more interesting was that the fire was about one mile from our task force operation and happened the very day our behavioral profiler told us we would have a fire within a week.

January 22, 2004. At 0311 hours, we had another turn with the arsonist. The fire was set in an apartment building, not a single-family home. The device was placed in an interior stairwell but not in such a way that it blocked people from getting out. The address was 3418 55th Avenue in the Bladensberg section of PGC. Interestingly, it was about one mile from the command post but in the opposite direction of the 83rd Street fire. Was the arsonist sending us a message?

February 6, 2004. We made a major presentation to the area police chiefs at the ATF lab in hopes of solidifying our resources and making sure everyone had bought into the case investigation. The weather was bad-snow and sleet. Schools opened and then closed. The entire area was in gridlock. We knew not many people would be attending-it was also a Friday. Boy, were we wrong! The room was filled. Everyone wanted to know about the case. We impressed many important people with what we had done. Now all we had to do was to catch him.

Later that same day, at 1456 hours, there was a fire in Fairfax County, Virginia. The fire took place at a garden apartment complex; a device was found in front of an apartment door directly next to the stairway. This one was blocking people’s egress. Luckily, the fire was in the middle of the day and while school was out. Not many people were in the building. Another twist: a fire in the middle of the day and no one is seen. The arsonist seemed to be stepping up his game.

February 14, 2004. This was Valentine’s Day. At 0456 hours, a fire took place in Montgomery County, Maryland, another case first. The fire was at 7700 Blair Road, a garden apartment. The device was placed in the stairwell between the first and second floors and blocked all passage from the upper stories to the exit. When the fire companies arrived, people were hanging from windows. Several had to be rescued. Four people were hurt; three, two of them children, had to run though the flames. During the evidence collection, a pair of burnt black pants was recovered. In an analysis conducted by the Montgomery County Crime Lab, DNA consistent with the DNA of the hair found at 4115 Anacostia Avenue in D.C. was found on the pants. The task force concluded that the DNA was that of the serial arsonist and welcomed Montgomery County police to the task force.

February 15, 2004. We flew in a sketch artist from the Houston (TX) Police Department so we could have a hand-drawn picture of our suspect. After interviewing one of the brothers from 4115 Anacostia Avenue in D.C., we had a new sketch of our arsonist.

February 18, 2004. We held another press conference to release the sketch and announce an increase in the reward money to $50,000. There were no more fires until April.

April 16, 2004. There was a fire at 2401 Rosecroft Village Circle East, the Oxon Hill section of PGC. All the parts of the device were present, and it was placed on the side of the house next to the front porch. Recovered in the debris was a burnt piece of advertisement for a local tax preparation firm. We traced the ad back to a local Giant Food store in Northwest D.C. We checked the store security tapes, but they had been taped over. We installed our own cameras in the store.

May 13, 2004. There was a fire at 3400 Beachcraft Drive in Fairfax County, about three blocks from the Cherry Arms Apartment complex fire. The fire occurred at 0438 hours at a single-family home. All the components of our device were present; it was placed at the front left side of the house.

August 30, 2004. A fire occurred at 4800 Elmwood Drive in Fairfax County. The device was placed directly in front of the entrance door. Luckily, the occupant had already left for work. There is a remarkable likeness to the surroundings of this home and that of the home on Beachcraft. Both homes were at the end of a block, both were backed up to woods, and both had a creek running beside them. There was a park nearby and new homes under construction in the rear. The homes were single story and had the same style landscape. Was our arsonist starting to look for a particular type of target?

September 8, 2004. At 0630 hours, there was a car fire at 2800 Channing Street, NE, D.C. The device was located under the driver’s side front wheel. The vehicle was owned by a Metro Transit Bus driver and was parked across the street from a transit bus barn. We had no cars in our case that we knew of to this point, but this fire caused us to wonder about how many fires our arsonist might have been setting that we did not know about. Our investigation had to be widened to look at all types of fires. How were we going to manage this? We would split our staffing and place them on a midnight shift rotation-one week of nights, one week of days.

September 20, 2004. At 0451 hours, we had a fire at 2804 30th Street, NE, D.C. The fire was one block from a fire at 2800 Evarts Street, which took place on June 5, 2003, before the task force was in operation. This fire took the life of Ms. Lou Edna Jones and was confirmed as the work of our arsonist. The fire was also two blocks from the Channing Street car fire 12 days before. Recovered at the fire was a piece of cloth wick, which was later determined to be consistent with an athletic sock. The lab reviewed the sock and extracted DNA found to match the DNA of our serial arsonist.

September 23, 2004. A fire occurred in Montgomery County at 10800 Amherst Avenue. The device was placed in front of a garage in a townhouse development. The arsonist was very busy and had us moving from one county in one state to another county in another state. The staff was focused but was getting very tired. Then, with no reasoning to support it, the fires stopped. We went 51 days without another fire.

December 5, 2004. A fire took place in Arlington County, Virginia. It was 0627 hours, and the fire was set on a reach deck of a single-family home. The Arlington Fire Department contacted the task force and discussed the case. The fire involved what investigators believed was a candle wrapped in a cloth and then soaked with gasoline. In the canvass of the scene, investigators located a Marine Corps dress cap and a pair of Marine Corps dress pants. These items were found on the block next to the fire. After talking with investigators, it was agreed that this did not sound like our fire, but since we had found a type of dress pants at the scene of one of our other fires, we wanted to be safe and collect all of the evidence Arlington had recovered and have it reviewed by the lab. We picked up all the evidence and pictures of the scene and took it to the lab.

December 6, 2004. A vehicle fire occurred at 5055 11th Street, NE, D.C., about two miles from the Channing, 30th, and Evarts locations. The owner was a Metro employee who used to work at the bus barn across from the Channing Street location. The device was found under the passenger side wheel in the front. During the review of the collected evidence, investigators found some writing on a black plastic bag: “Made in China for the Cornelius Sho…” Based on this, investigators spent long hours on the Internet and located a Cornelius Shoppe in England. Through a collateral request to the United States Secret Service, arrangements were made to interview the storeowner.

December 7, 2004. A fire took place at 4205 53rd Avenue in the Bladensburg section of PGC. The device was located in the stairwell between the first and second floor, blocking all passage from the upper floors. We got lucky; nobody was hurt.

December 10, 2004. We had a fire at 3021 Yost Place, NE, D.C. and found a device on the front porch directly next to the front door. During our evidence-collection process, we were able to identify a black bag similar to the one recovered at the December 4th fire. This time, investigators were able to piece the entire thing together, “Made in China for the Cornelius Shopping Bag Company.” We are able to determine that this bag company was located in Richmond, Virginia, and dispatched two investigators to drive the 100-plus miles right away. They found that the bags were distributed to just two stores in the Maryland, D.C., Virginia region. We identified the stores as Circle Seven stores, owned by the same person. One was located at 2713 Martin Luther King Avenue, SE, D.C.; one was located at 740 Kenilworth Avenue, NE, D.C. Investigators quickly contacted the stores’ owner and solicited his cooperation. Video surveillance was established in both stores. It was our theory that the arsonist would return to one of these two stores to purchase additional components for his device. When he returned, we would be waiting.


Task force members worked for several days to devise plans to catch the arsonist. We developed what became known as the “Black Bag Operation.” We provided each store with bags that had a small stainless steel tag in the bottom of each bag. These tags were numbered and set up with high numbers for one store and low numbers for the other. We arranged with the stores to place in these bags only purchased one-gallon jugs. The prevailing thought was that when there was a purchase of one of our bags, it would be used as part of a device. The stainless steel chip would remain after the fire. We used metal detectors to locate the chip and identify the store from which it came. To better identify the store, we made daily trips to the stores at 6 a.m. and picked up a bag. This way, we were able to keep track of the numbers of items sold based on the number of bags used. Each morning at our 8 a.m. briefing, investigators spent at least one hour setting up bags. This was not rocket science, but the hard work of dedicated investigators. We were now in the year 2005 and still had not caught the arsonist.

February 10, 2005. We brought in the boys from 4115 Anacostia Avenue in D.C. and had two of them undergo clinical forensic hypnosis by a doctor and an ATF forensic sketch artist. The information gained in the process was to be used in a future plea to the public for their help.

We moved from February to March and from March to April with no additional fires. In fact, we had not had a fire since December 10, 2004. The same thoughts came up about our arsonist: Was he dead? Was he in jail? Did he go on vacation? Does his job take him out of the area? We continued to work our overnight schedule and cover all three areas as best as we could.

April 1, 2005. The task force was notified that the Marine Corps dress pants found at the Arlington fire had DNA that matched the DNA found on the other items related to the serial arson suspect. The Arlington fire was ours. With this information, the task force determined that our suspect was a current or former member of the Marine Corps or had strong ties to the Corps.


A meeting was held with attorneys and members of the Naval Criminal Investigative Services. We wanted to access the Armed Forces DNA database in an effort to match our suspect’s DNA profile with a member of the military. What we found was that the military stores only the raw blood taken from each member and uses it only for identification purposes at a later date. The DNA profile is not placed into a database for query purposes. Although we could not gain the information we had sought, NCIS investigators made us aware that they knew of two arson suspects they believed had been setting cars on fire almost three years ago.

Interestingly enough, one of those suspects lived very near to one of our target stores on Martin Luther King Avenue. We immediately set up surveillance on this suspect, identified as Thomas Anthony Sweatt, a black male born on 11/1/1954. We watched his home, which was located at 556 Labaum Street, NE, D.C., and his place of work, a Kentucky Fried Chicken Store at the intersection of New York Avenue and Bladensberg Road, NE, D.C. This location is just blocks from the Evarts fire, the Channing Street fire, the Yost Street fire, and others on our list.

His photograph was shown to the witnesses from 4115 Anacostia Avenue; none provided a positive response. We found nothing to indicate this could be our guy: He had no major criminal record, he worked for KFC for 20 years, he had a quiet home life, and he worked in the church. But, we stayed the course. He would be ruled in or out just as any other person brought to our attention.

April 19, 2005. Task force investigators interviewed Sweatt at his place of business. When asked if he was the serial arsonist, he replied, “Why would I want to burn those beautiful homes when I am trying to be a homeowner myself?” He did not say no. Investigators did not think this was the guy and asked if he would be willing to provide them with a sample of his DNA. He agreed. A sample was sent directly to the lab. We maintained static surveillance on Sweatt over the weekend and asked the lab to expedite the DNA check so we could move on.

April 25, 2005. At 0900 hours, the task force was notified by the Montgomery County Police Department Crime Lab that Sweatt’s DNA was an exact match with the DNA recovered at 4115 Anacostia Ave, NE, D.C.; 7700 Blair Road, Montgomery County, Maryland; 2804 30th Street, NE, D.C.; and the Marine Corps dress pants from 301 North Bryan Street, Arlington, Virginia.

All task force members met and initiated a gag order to attempt to prevent any leaks. Knowing we did not have much time before the word got out, the special agent in charge of the Baltimore Field Office and the head of the task force were asked to notify all the senior staff within all the participating agencies of the pending arrest. The command post went to 24-hour operations. Teams were set up for 24-hour surveillance on Sweatt until the arrest, because of public safety concerns. Meetings were held with U.S. attorneys to discuss the charges.

April 27, 2005. Sweatt was arrested at a KFC on Marlboro Pike in the District Heights section of PGC. He was then transported to a location for interview. He was interviewed for one hour and 45 minutes; he confessed to being the MD/D.C./VA serial arsonist. A video camera was introduced at this time to more effectively document the confession. The interview continued for another four hours. Sweatt confessed to setting fire to four apartment buildings, six vehicles, 39 residences, and one attempted arson.

During the execution of the search warrant on his home, investigators found 75 videotapes, additional Marine Corps dress pants, a Navy hat, and a newspaper article of a fatal fire on Montello Street in Washington, D.C. in 2002. A review of the recovered tapes revealed a number of news stories referencing fires set by the serial arsonist over the past two years. They also revealed a number of military personnel in uniform and Metro Transit drivers in uniform unknowingly being recorded.

On May 10, 2005, just 14 days after his arrest and confession, Sweatt signed a plea agreement, pleading guilty to five counts of arson of buildings used in Interstate commerce, two counts of the use of a firearm during a crime of violence, six counts of possession of an unregistered firearm, one count of first-degree murder while armed, and one count of second-degree murder while armed.

The task force also debriefed Sweatt on six additional occasions since his plea of guilty in formal proffer sessions and informal ride-alongs in the affected three states. During this time, Sweatt provided investigators with information on more than 303 fires for which he was responsible. Sweatt admitted he had been setting fires in the metro region since 1978 and that those fires to date totaled to 353. Sweatt continues to be cooperative with the task force so cases could be closed.

On September 12, 2005, Sweatt was sentenced to two life terms, plus an additional 136 years.

ROBERT M. LUCKETT, a 30-plus-year veteran of the fire service, is an assistant fire marshal for the Alexandria (VA) Fire Department. He is a certified fire investigator by the state of Virginia and also is certified as a law enforcement officer by the Virginia Department of Criminal Justice and the Virginia Fire Marshal Academy.

No posts to display