September 1, 2018 Vol. 1, No. 3

Sifting the Ashes

Advanced” Students Need to Know the Basics

So, I was looking at my transcript and noticed that I had not taken some of the really basic modules. I recently taught at the ATF’s Advanced Origin and Cause/Courtroom Testimony Course (AOCCT) on subjects that were anything but “advanced” (Fire Chemistry and Fire Dynamics) and I wanted to see how well my course conformed to the module in Fire Chemistry.

As I moved through the module, I found that I appreciated the way it was structured, and how the information flowed. After completing the module, I looked at the recommended reading list and found my Chapters 2 and 3 there. No wonder the module made such good sense!

There was some pushback from one of the attendees at the course mentioned above, who wondered how knowing that a cubic foot of propane requires 25 cubic feet of air to burn completely had ever helped in an origin and cause determination. I was able to describe a case with a gas water heater that caused a fire because it was starved for air, but the number 25 was not necessary—just the knowledge that the gas flame required an adequate supply of air for combustion and ventilation.

The larger answer is that if we are to engage in a professional pursuit, we need to know the basics of what makes fires burn, and how the molecules interact in the process of combustion. Without an appreciation for what makes things happen, “reading” fire patterns involves no more skill than astrology. Professionals need to know the terminology of their profession, the units used to measure energy, power and heat flux, i.e., the basic knowledge of fire chemistry and fire dynamics that are required by NFPA 1033.

Would you want to consult with a doctor who did not know basic anatomy and physiology, or who did not know the normal limits for blood pressure? Would you hire a contractor who did not know the building code, or who did not know the appropriate spacing of floor joists? Or how many square feet there are in a square yard? So how is it that there are fire investigators, people with many years of experience, and letters behind their name, who can’t tell you what the basic units of energy are, who can’t tell you what heat flux is?

I look forward to the day when a roomful of students at an “advanced” course already meet the minimum criteria in NFPA 1033, and only need instruction on these subjects as a refresher.

For an overview of how qualifications challenges are playing out in actual cases, readers are invited to review the three publications from 2017 on the website.

Case Study of the Month

State of Michigan v. David Lee Gavitt [1]

David Gavitt and his wife Angela were in bed at 10:30 PM on the night of March 9, 1985, when they were awakened by their dog and became aware of a fire in the living room. Angela ran down the hall, and David ran outside to attempt entry into the children's bedrooms through the windows. He cut himself badly in the process and was hospitalized for burns and severe lacerations.

By the next morning, investigators from the Michigan State Police (MSP) had concluded that the “distinctive flammable liquid burn patterns” and “multiple origins” in the house indicated that the fire had been set with ignitable liquids, and David was arrested.

The investigators collected 6 samples of debris on March 10, all of which tested negative. They returned to collect more samples on March 20 from the front yard where carpet had been thrown out during the initial scene excavation. The samples were submitted to the MSP laboratory where they were analyzed by an individual who had experience as a food chemist, but no serious experience as a fire debris analyst. He found that two of the samples, both from the front yard, were positive for “residues of highly evaporated gasoline.”

The chemist ran an additional “flame spread test” involving an attempt to ignite the carpet with a Bunsen burner. Three times he applied the flame to the carpet and three times the fire went out shortly after removal of the flame. In a fourth test, he poured a small amount of gasoline on the carpet and lit it with a match. Unsurprisingly, the carpet stayed lit when the blowtorch was removed. This was “proof” that the carpet would not have burned had it not been for the application of gasoline.

The photographs of the scene clearly show that origin was in the living room, Figure 9.8 (a), but nothing in the photos suggested multiple origins.

Figure 9.8 (a) Two photos of the fire damage in the Gavitt living room.

In September 2010, the author was contacted by Michael McKenzie and asked to review the case. I told him that if gasoline really was present in the living room, there was not much that I could do. I was provided with the GC-FID chromatograms and was appalled by what I saw. The two allegedly positive samples matched neither each other nor the gasoline standard, and the quality of the work was unbelievably poor.1 Figure 9.8 (b) shows the comparison of the allegedly positive samples and the standard, and Figure 9.8 (c) shows a publication quality chromatogram from 1977. The difference in resolution is quite remarkable.

Having determined that the chemistry was not credible, I continued my analysis of the fire scene investigation, and found it to be equally unreliable. A timeline analysis revealed that the living room burned in a fully involved condition for an extended period of time, in excess of 13 minutes, so reading fire patterns was not reliable, although that was not known to the investigators of 1985.

Figure 9.8 (b) GC-FID chromatograms from two samples incorrectly called positive for gasoline. In each chart, the sample chromatogram is gold and the gasoline “standard” is black. The samples match neither gasoline nor each other.

Figure 9.8 (c) Figure 9.8 (c) Publication quality GC-FID chromatogram of gasoline from 1977.

I handed off the chemical analysis to a colleague, Craig Balliet of Barker and Herbert Analytical Laboratories, who agreed that the chemistry was abominable and prepared an affidavit to that effect. I prepared an affidavit about the fire scene analysis, and both were submitted to the court in support of a motion for relief from judgment.

The prosecutor attempted to have the chemistry re-evaluated by the MSP laboratory. They reported, “Because the original sample processing methods are unclear and the instrumentation that was used is unavailable for evaluation of the reference samples, no definitive conclusion could be made regarding the presence or the absence of gasoline in the questioned carpet samples which were previously reported to contain highly evaporated gasoline.” The ATF laboratory declined to re-evaluate the data, stating that they had no protocols for doing so, having abandoned GC-FID 15 years ago. They also stated that because the data had already been reviewed by multiple experts, it was unlikely they would have anything to add. The prosecutor then turned to Dirk Hedglin of Great Lakes Analytical Laboratories. Mr. Hedglin agreed with the analysis conducted by myself and Mr. Balliet.

The prosecutor also reached out to ATF Special Agent Michael Marquardt to have him review the fire scene analysis. Mr. Marquardt reached the conclusion that there were “insufficient data (facts, circumstances, information, evidence, etc.) under current fire investigation accepted practices, procedures, and standards to clearly support an incendiary fire cause classification to a reasonable degree of scientific certainty.” (NOTE: “Scientific certainty” does not exist, but that is a subject for another time.)

The prosecutor delayed for an additional eight months after being told that the evidence against Mr. Gavitt was not supported by science, arguing inter alia, that although fire investigation practices had changed dramatically, Mr. Gavitt could have contested the faulty GC-FID analysis in his original trial. Despite the prosecutor’s misgivings, he agreed not to object to a new trial. The prosecutor continued to maintain that he believed Mr. Gavitt was guilty, even though there was no evidence that indicated arson, and throughout the 27-year pendency of this case, nobody had even suggested what Mr. Gavitt’s motive might have been. Because the evidence would not support a guilty verdict, on June 26, 2012, Mr. Gavitt was released from custody. He is shown, accompanied by his lawyers in Figure 9.8 (d). Mr. Gavitt's name has since been added to the National Registry of Exonerations. As is frequently the case with exonerations, no determination was ever made as to the actual cause of the fire. Seeing what they thought was evidence of arson, the investigators did not collect any data that would have indicated an accidental cause.

Figure 9.8 (d) David Gavitt leaves the penitentiary after serving 27 years for a crime that nobody committed. He is accompanied by his attorneys Imran Syed and Michael McKenzie.

The Gavitt case formed the basis for a paper on the subject of what should happen when the science underlying a conviction changes. Imran Syed and Caitlin Plummer of the University

of Michigan Law School’s Innocence Clinic authored a paper called “Shifted Science and Post-Conviction Relief,” which was published in 2012 [2]. Error analysis

The original fire investigators misinterpreted critical data by stating that the low burns were indicative of an incendiary cause. (Not at all unusual for 1985). Further, they engaged in two-dimensional thinking by stating that the lack of communication at the floor level indicated multiple origins.

The chemist (who has since died) was not properly trained or qualified in fire debris analysis, and the MSP was negligent in letting him conduct the analyses. He misinterpreted critical data, by concluding that the chromatograms indicated the presence of gasoline when they clearly did not. The chemist’s flammability test was beyond the pale, and constituted irrelevant data. This chemist had no experience in conducting such tests, and grossly overstated the meaning of his test with and without gasoline.

These errors cost David Gavitt 27 years of his life. Significance

David Gavitt’s wrongful conviction was a typical outcome of fire investigation in the 1980s. His case has encouraged discussion of how the law should respond when the science underlying a conviction changes. After his release, Mr. Gavitt testified before the Michigan State Legislature in support of the Wrongful Imprisonment Compensation Act, which went into effect in 2017. Despite being one of the main sources of inspiration for this legislation, The Michigan Attorney General’s Office has resisted paying any compensation to Mr. Gavitt, insisting that he do more to “prove his innocence.” That can be a tall order in a case where no crime was ever committed, but at least the good citizens of Michigan can rest assured that their criminal justice professionals are continuing their mission of afflicting an innocent man. For a more detailed description of this continuing miscarriage of justice, see the coverage at this link: The crafting of the legislation was intended to prevent individuals who got off “on a technicality” from collecting compensation, but the AG’s interpretation of the law seems to be that this applies to every exonerated individual. Clearly, with respect to Mr. Gavitt, the new law failed to accomplish the legislature’s goal.


1 A colleague presented these chromatograms to an audience of forensic scientists at the Bureau of Criminal Apprehension Laboratory in Minnesota in 2012. She reported that there was an audible gasp when the scientists saw these charts.

2 Plummer, C. and Syed, I., (2012), “Shifted Science” and Post-Conviction Relief , 8 STAN . J. C.R. & C.L. 259.

If you are not on the mailing list, please click on “Join the Mailing List” to automatically receive the monthly newsletter.

Thank you for your interest. If you have any suggestions for future articles please email me at

firstname :