Sifting the Ashes
July 1, 2019 Vol. 2, No. 7

Comparison Samples: Collect Them If Possible

The subject of comparison samples collected to help the laboratory determine whether the ignitable liquid residues (ILR) they detect are foreign to the background has never been far from the forefront of discussions about fire debris analysis and evidence collection. NFPA 921 recommends collection of comparison samples and has done so since its first edition.

There is new draft standard under consideration for adoption by ASTM Committee E 30 on Forensic Sciences, entitled “Standard Guide for a Systematic Approach to the Analysis of Ignitable Liquids. While the draft standard encourages the collection of comparison samples, it contains a statement that needs to be read very carefully and understood in the context of laboratory analysis, not fire investigation.

The offending sentence states, “The submission of a comparison or control sample is not a necessary requirement for a valid ignitable liquid determination.” This statement is true most of the time, but there are many times when although the chemical analysis might be “valid,” the result does not represent relevant and reliable evidence.

Consider the case of a hardwood floor that was finished with polyurethane 20 years before the fire. Due to the amazing sensitivity of laboratory separation and analysis methods, it is possible to detect the medium petroleum distillate solvent that was laid down and trapped in the polyurethane all those years ago. How does finding that residue help with your evaluation of the fire scene?

In the absence of a comparison sample which tests negative, that finding of MPD is meaningless but potentially very misleading. Over the course of my career, I have seen numerous cases where the submission of a comparison sample would have prevented a wrongful accusation of arson.

So don’t let the statement in this new document mislead you. The proponents of this statement make a distinction between an “ignitable liquid” determination and an “accelerants” determination. This distinction may be lost on most fire investigators.

In fact, calling it an ignitable liquid determination rather than an ignitable liquid residue determination confuses things even further. If there was no foreign ignitable liquid on the floor, the finding of the solvent trapped in the surface of the floor is of no value in helping you understand the cause of the fire.

The bottom line for comparison samples can be stated in four words: Collect them if possible. If your analysis of the cause of the fire relies on a laboratory report, without the benefit of a comparison sample, your analysis may be in trouble.

One more reason to collect comparison samples is that extinguishing water can sometimes spread ignitable liquids from an area where they belong to an area where they don't belong. For a detailed analysis of this problem, refer to the article entitled “The Possibility of Ignitable Liquid Contamination in Flooded Compartments” by John Black, Justin Geiman and Ray Kuk published in the Fire and Arson Investigator in 2016.1 This article is available for free to IAAI members. Non-members who need a copy should contact me via e-mail (

This should go without saying …

Unfortunately, it can't. One of the dumbest things an expert witness of any kind can do is to misrepresent their credentials. In fact, the International Association of Arson Investigators Code of Ethics used to have a specific prohibition of this practice. The words were, “I will make no claim to professional qualifications which I do not possess." That was one of 10 rules.

The current Code of Ethics has six rules, and I suppose that misrepresentation of credentials would fall under the rule requiring us to "Conduct both our personal and official lives so as to inspire the confidence of the public."

Despite not having a specific reference to misrepresenting credentials in the Code itself, the SOP for the Ethical Practice and Grievance Committee states that the basis for a complaint may include one of seven categories, number 5 of which is "Misuse or misrepresentation of an IAAI professional designation or certification.” The full SOP is available at

The headline on this sad story about an IAAI member testifying at a murder trial in Syracuse New York says just about everything that needs to be said.

But here is how the clever investigator defended himself.

1) He is certified by the state of New York

2) He is a member of the IAAI.

Therefore, he can put “IAAI-CFI” behind his name. He did not even have the required four years of experience to challenge the CFI exam.

Ultimately, while the investigator admitted to misrepresenting his credentials, he stated he did not intend to do anything fraudulent, and the evidence against the defendant was sufficient that the jury convicted him despite the unethical conduct of the fire investigator.

It is headlines and stories like this that encourage people to take what is written on the resume with a grain of salt. Most of the time, checking credentials is not at all difficult, and it is becoming a component of due diligence when litigation is involved. Whenever I am asked to review a case, I always review the investigator’s CV.

Very few people are misguided enough to “fudge” (a euphemism for lie on) their resumes, but puffing is pretty common. I see this frequently when reviewing a CV where the investigator claims (legitimately) to be certified by the IAAI. The IAAI certification program is, in turn, accredited by the NFPA Board on Fire Service Professional Qualifications (Pro Board) and some investigators include Pro Board certification on their resumes. I have seen others state that they are “NFPA 1033 certified.” Neither the Pro Board nor the NFPA certifies individuals, though you can get a pretty piece of paper like the one below.

Figure 1. I had already been certified for 25 years when I got the above certificate by sending in a $15 fee. I never listed this certificate on my resume, and let it expire when my CFI was renewed in 2016.

While it is true that any IAAI certificate holder can obtain for a $15 fee a certificate stating “It is hereby confirmed that [the certificate holder] having been examined by an accredited agency in the National Professional Qualifications System is certified as Fire Investigator,” only someone knowledgeable in the business would understand that the IAAI-CFI and the Pro Board certificate represents one, not two certifications. In fact at the bottom of the Pro Board certificate, it states "Certification issued by International Association of Arson Investigators, Inc."

Putting your Pro Board certificate on your resume makes it look like you are trying to claim two certifications when in fact you have one. It just shouldn’t be done. If you do it, you will likely face questions on cross-examination when you attempt to qualify as an expert. Rather than enhancing your credibility, it will diminish it.

From Frye to Daubert to Frye and Back to Daubert

What's the difference between a Frye state and a Daubert state? That question has certainly been on the minds of litigators in the great state of Florida recently.

While the qualifications of an expert witness are subject to challenge whether the state follows Daubert or Frye, challenges to expert opinions in the form of motions in limine are more common in Daubert states. Frye has one criterion: is the methodology “generally accepted in the relevant community?” If so, the opinion comes in. While “general acceptance” is one of the factors that a judge can take into account under Daubert, there are also other factors, such as whether the methodology on which the testimony is based is centered on a testable hypothesis, the known or potential error rate, the existence and maintenance of standards, and whether the methodology has been peer-reviewed. In a nutshell, challenges are more likely in a Daubert state than in a Frye state. You can see which state yours in by referring to this article:

The Florida legislature in 2013 passed a law stating that Daubert would be the rule. Then in late 2018, the Florida Supreme Court stepped in and said, “No, Frye is the rule.” Recent changes to the Florida Supreme Court membership, making it more conservative and more protective of business interests resulted in them going back in 2019 and stating that we are going to be a Daubert state. For deep a more detailed analysis of the Florida changes, see the post by Cozen O'Connor attorneys Richard Maleski and Joseph Rich.

New Washington Post Series on Fixing Forensics

Radley Balko, the Washington Post’s criminal justice reporter (and a friend) has started a series in The Washington Post on the problems with forensic science and how to fix them. I was honored to be asked to participate. Here is the first installment:

Case Study of the Month

This month's case study shows just how far we have come, as a profession. The case described goes back to 1981, when everything was done the old fashioned way.

Ray Girdler was convicted of two counts of murder and sentenced to two terms of life without parole in 1983. In 1991, after an extraordinary evidentiary hearing, the judge overturned the conviction that happened in his courtroom, and with the consent of the State’s Attorney dismiss the charges.

While in prison, Girdler had the lucky break of meeting John Henry Knapp, who was also convicted of arson and homicide that resulted in the death of his children. Knapp, who was sentenced to death, had obtained the services of Larry Hammond, a criminal defense lawyer from Phoenix, and Girdler got in touch with Hammond to tell him he had pretty much the same problem as Knapp did.

Knapp’s story ended up in a book called Triple Jeopardy by Roger Parloff.

Ray Girdler’s story ended up in my book.

Many of the documents referred to in this case study are available at this link.

State of Arizona v. Ray Girdler i

Ray Girdler lived with his wife and daughter in a mobile home near Prescott, Arizona. At about 2:45 a.m. on a cold November night in 1981, Girdler was sleeping in the bedroom when his cat’s meowing awakened him. When he stood up, the room was full of smoke. He woke his wife, then ran down the hall and saw fire in the living room. He ran to the car to get a fire extinguisher, and his wife ran to the middle bedroom where their 2-year old daughter was sleeping. A floor plan of the Girdler residence is shown in Figure 1. When Girdler returned to the mobile home, the fire had grown to such an extent that he could not reenter. This was a 1964 model mobile home, and the interior finish was thin plywood paneling (flash paneling). Girdler was unable to save his family but managed to escape, barefoot and in his underwear, and ran to a neighbor’s house to call the fire department. By the time the fire department arrived, the roof over the center of the home had collapsed.

Girdler’s neighbor gave him some clothing to wear and a cup of coffee. When one of the firefighters went to the neighbor’s house to interview Ray and other witnesses, he saw Girdler fully dressed in the middle of the night and immediately became suspicious. His report stated, “…the first thing I noticed was that he was dressed. He had on a T-shirt, pants, and in front of the couch next to him was a pair of socks and tennis shoes. He was drinking a cup of coffee and smoking a cigarette.” It apparently did not occur to the firefighter that Ray had not yet put the socks and shoes on—he simply assumed that he had taken them off.

The local fire marshal came out and was able to discern “multiple points of origin” in the trailer, shown in Figures 2 and 3. This “determination” was made by 4:00 a.m., while it was still dark. The Arizona State Fire Marshal was called, and they sent out a relatively inexperienced investigator, who also “saw” evidence of an intentionally set fire: low burning, irregular patterns, holes in the floor, and crazed glass. Later, a senior Arizona Deputy State Fire Marshal claimed to have reviewed the evidence, but in fact only talked to his subordinate by telephone. Three days after the fire, before seeing the scene, or even photos of the scene, before getting the (negative) results of the laboratory analysis of fire debris, and before the autopsy report was written, this supervisor told the District Attorney that he was “absolutely certain” that this was an arson/homicide fire. Girdler was arrested the next day.

Figure 1. Floor plan of the Girdler mobile home.

Figure 2. Exterior view of the Girdler mobile home. A fire marshal had discerned multiple areas of origin before dawn broke on the morning after this fire. (Courtesy of David M. Smith, CFI, Associated Fire Consultants, Bisbee, AZ.)

Girdler’s wife and daughter had relatively high concentrations of carboxyhemoglobin in their blood (74% and 87%, respectively), and the county medical examiner opined that this was further proof that the fire was accelerated.2 Given that Ray was in the house, he was the obvious suspect. If this was a set fire, Ray set it. This is the case with many of the false accusations the author has seen, and the main reason that a first hypothesis of accidental cause is warranted.

Girdler was brought to trial in front of Judge James Sult in June of 1982 and was convicted of arson and two counts of homicide. During the trial, the fire marshal, who testified that he determined arson in 80% of the fires he investigated, repeatedly stated that nothing other than the presence of flammable liquids could account for the condition of the mobile home. About a hole burned in front of the sofa, he stated, “And this is an area without any possible natural or accidental fire cause.” He further testified that it appeared that Ray had broken the leg off a coffee table and beaten his wife and daughter over the head with it. The injuries to the victims’ skulls were actually thermal injuries, but that possibility was not considered by the fire marshal, nor was it presented to the jury. The earliest fire reports describing the condition of the victims indicated “possible skull fractures.”

In a remarkable example of the wide latitude afforded to experts, the fire marshal was allowed to opine that the crazed glass shown in Figure 4 had tremendous significance. Some of it was crazed, and some was not. The fire marshal’s report stated:

Much of this glass was heavily heat-crazed and free of smoke deposit, indicating a rapid buildup of intense heat and close proximity to initial fire, which is consistent with the expected results of a liquid accelerant fire. A significant volume of the arcadia door glass was found which was free of both heat and smoke damage in a configuration indicating breakage by physical force other than explosive force, prior to or in the very early stages of the fire. ii

Figure 3. Another view of the exterior of the Girdler mobile home. (Courtesy of David M. Smith.)

The fire marshal elaborated on what he meant by “physical force” at the trial. With no objection from the defense, he testified, “I would say it was one of two things. Either there was a struggle or there was someone in there with a fit of rage.”

During his trial testimony, the fire marshal stated that the wife and daughter “never had a chance” and were immediately overcome by this rapidly moving fire. After the conviction, there was a penalty phase, during which that same fire marshal testified that Mrs. Girdler and her daughter had died a “slow, painful death,” suffered “excruciating pain,” and had their lungs “severely seared.” Actually, the autopsy found no evidence of damage to the trachea or lungs. These victims inhaled a few breaths of smoke that contained a high concentration of CO.3

Figure 4. Crazed and non-crazed glass outside the Girdler mobile home. The fire marshal testified that the explanation for the condition of the glass was that the defendant broke out the window “in a fit of rage” and then set the fire. Heat from the accelerant caused the glass to craze but the previously broken out glass was protected. (Courtesy of David M. Smith.)

In August 1990, in an extraordinary hearing that went on for several days in front of the same trial judge that oversaw the first trial, Ray Girdler, with the assistance of defense attorney Larry Hammond, was granted a new trial. The conviction was overturned and the sentence was vacated on the basis that “new evidence,” in the form of new knowledge about the phenomenon of flashover, made the testimony that the jury heard about flammable liquids false.

Further, a respected medical examiner testified that the proposition that the high COHb (carboxyhemoglobin) levels found in the fire victims could only result from an accelerated fire had no scientific basis. The original medical examiner remained adamant but, under cross-examination by Hammond, had to admit that it was only because “it stands to reason” that he held his beliefs about accelerated fires.

The fire marshal was confronted about the possibility that the burn patterns he attributed to accelerants could have been caused by radiation, although his trial testimony repeatedly stated that accelerants were the only possible explanation. At the hearing, he admitted that alternative explanations existed that he had not disclosed to the judge or jury at the trial. He was allowed to “explain” his reason for not disclosing such alternatives. His answer, reflected in the record, was that he did not see his function as requiring him to “discuss every possibility that I ever considered.” Among other findings, the judge ruled, “The fire in the Girdler mobile home could reasonably be attributed to noncriminal causes. Specifically, scientific evidence now available supports the finding that the fire may have been a “flashover” fire ignited by accidental or other noncriminal means and not involving the use of flammable liquids or accelerants.”

A new trial was ordered, but the judge called a temporary halt in the proceedings. Judge Sult was thoroughly chastened by hearing this evidence and overturned the first verdict on his own initiative, rather than requiring Hammond to appeal to a higher court.4 Girdler had been Judge Sult’s first capital case, and he had been particularly harsh in sentencing. In a scathing rebuke, the judge had told Girdler that he was going to impose consecutive life sentences, making him eligible for parole at age 100. He stated:

Jennifer Ann Girdler was a baby of the age of two years, a totally innocent and helpless victim, completely dependent upon you not only for the necessities of life but for life itself. Your killing of her constitutes the most vile betrayal of the highest trust which is placed on a human being. …

It is mandated in our law that the Court consider in every sentencing the need for establishing a deterrent to others like you who will inevitably consider the same method you adopted to rid yourself of your problems. Such deterrence can only occur if in this case you are required to spend the maximum amount of time possible imprisoned for your deed.

Larry Hammond had retained David M. Smith to assist in the defense of the second trial, and the State sought assistance from Barker Davie, in addition to using its own “in-house” experts. Presented with dueling experts, the judge ordered the parties to identify a third expert to review the evidence for the benefit of the court. Both Smith and Davie agreed to recommend this author as a neutral, qualified, independent expert and I accepted the court appointment.

I was provided with all the expert testimony from both the trial and penalty phase, as well as some post-conviction hearing transcripts. Additionally, all the 8” × 10” photographs used in the original trial were provided, as was some of the actual physical evidence, including the crazed glass. As with most of these cases, laboratory analysis failed to identify the presence of any ignitable liquid residue in any of the samples collected from the scene.

My charge from Judge Sult was to help him understand why Girdler had been convicted the first time, and whether he would be convicted if tried again. The answer to the first question was not difficult to determine. The testimony of the fire marshal had been devastating, although it was internally contradictory and based on mythology and rank speculation. The alleged pour pattern was nothing more than damage to a vinyl floor caused by radiant heat. Numerous boxes were stored in the hallway, and the edges of the alleged pour pattern were straight lines with right angles in them, shown in Figure 5. Even the irregular pattern on the bedroom floor shown in Figure 6 was clearly the result of protection of various unburned areas. There was nothing even vaguely liquid-appearing about most of the alleged pour patterns, other than that they were on the floor and liquids lay on the floor if they are spilled.

Crazed glass results from the application of liquid to hot glass, not rapid heating, as has been repeatedly discussed. Applying the flame from a propane blowtorch to a piece of the glass from the Girdler residence did not result in crazing, just thermal cracking. Applying a few drops of water to the hot glass did result in crazing. That the fire marshal testified that crazing indicated the presence of accelerants near the glass was excusable given the state of the art in 1982, but his further speculation that Girdler had broken the glass in a struggle or a fit of rage was pure fantasy. Nonetheless, it was allowed into evidence without objection from defense counsel and the jury gave this ridiculous testimony the special weight that juries give to expert testimony.

Figure 5. Alleged “pour pattern” in the hallway of the Girdler mobile home. This floor would have presented a pattern of uniform charring were it not for the presence of boxes on the floor. (Courtesy of David M. Smith.)

Figure 6. Continuation of the alleged pour pattern into the bedroom. The edges of the pattern are more irregular here than in the hallway because the objects providing the protection to the unburned areas were not boxes, but items of clothing and bedding. (Courtesy of David M. Smith.)

I provided Judge Sult with a 20-page, single-spaced report, outlining all the myths and misconceptions testified to by the fire marshal. My report also stated that if Girdler were to be tried again, Mr. Smith would certainly point out all these errors, assuming the judge decided to let the fire marshal put the same rubbish in front of the jury again. The prosecutor agreed to dismiss the case.

Ray Girdler had served more than 8 years in the penitentiary and brought suit for malicious prosecution against the fire marshal, a suit that was eventually settled.

Error Analysis

The first fire marshal was thinking in only two dimensions when he perceived “multiple origins” in this mobile home that no longer had a roof. The supervisor, once he actually looked at the evidence, misinterpreted several irrelevant artifacts, particularly the “pour patterns” on the floor and the crazed glass. He also misinterpreted the thermal injuries to the skulls of Mrs. Girdler and her daughter as resulting from physical trauma. His failure to read the autopsy can be most kindly described as overlooking critical evidence, but is more accurately characterized as dereliction of duty.

The medical examiner, in providing support to the unsupportable musings of the fire marshal, misinterpreted critical data by assigning significance to the carboxyhemoglobin readings where none existed.

Even the 1990 review of the fire patterns was flawed, when they were again identified as “pour patterns,” although Davie, to his credit, published photos of the fire patterns and requested feedback from the fire investigation community on their significance, or lack thereof. iii


The Girdler case was one of the first in which an arson conviction was reversed based on the new understanding of flashover. Critics of the decision called it “the flashover defense,” but reasonable fire investigators began to take the phenomenon seriously. Newsweek and the Los Angeles Times published Girdler’s story and brought the relatively new fire science to the public’s attention.

References See also

1  Black, J., Geiman, J., and Kuk, R. (2016) The Possibility of Ignitable Liquid Contamination in Flooded CompartmentsFire and Arson Investigator, 67(1):26–31.

2 All this really proves is that the fire was ventilation controlled, and therefore the products of combustion contained high levels (1-10%) of CO. See NFPA 921 at 25.2.1. There is an

discussion of carbon monoxide production and transport in the SFPE Handbook.

3 It would be revealed eight years later that the fire marshal had testified at the first trial without the benefit of having actually read the autopsy reports.

4 Because the judge overturned himself, there is no appellate record of the grant of a new trial.

i State of Arizona, Plaintiff, v. Ray Girdler, Jr., Defendant, in the Superior Court of the State of Arizona in and for the County of Yavapai, No. 9809.

ii Dale, D., (1981) Office of the State Fire Marshal, Report of Fire, Mobile Home Fire with Two Fatalities, DR #81-05, Phoenix, AZ.

iii Davie, B., (1993) Flashover, National Fire and Arson Report, 11 (1), 1993, 1.

Subscribe to Our Newsletter

firstname :