|Terri's Fire Investigative
"Cursory official investigation, ... junk science ... perfunctory dismissal of wire arc ignition."
An investigative report written for arson defendant Terri Hinson Strickland by Gerald L. Hurst, consulting chemist, of Austin, TX.
March 16, 1998
Re: Terri Hinson case
Dear Mr. Wood,
Following your request, I have made a thorough study of all the documents you submitted to me with the object of evaluating the physical evidence and determining the scientific credibility of the conclusion drawn by the prosecution and insurance compa ny investigators based on that evidence. I have submitted this material to other well-qualified experts for their critique to make sure that I have not included any statements that will not themselves stand up to careful scrutiny by readers knowledgeable in the areas covered and familiar with the state of the art in the relevant fields. If any of my work is unclear in meaning from the standpoint of the lay reader, please let me know in order that I may correct the language to make it more lucid.
Negative corpus fire evidence. This is a negative corpus case relying solely on presumed burn pattern evidence, that is, the ability of the investigator to arrive at a meaningful conclusion of arson by the process of eliminating "all possible accidental and natural causes" for the fire patterns that are left in the wake of a structure fire.
Negative corpus evidence is considered an unacceptable evidentiary basis for arson. When the ignition source cannot be ascertained, the proper classification of the incident is "undetermined." The investigator may seek other evidence to prove a hunch or personal theory but it is unacceptable to classify the fire as "suspicious," or worse, as "arson."
NFPA 921, the standard for fire investigation. The most widely accepted authority on acceptable arson investigation practice is the National Fire Protection Association code NFPA 921 which describes the proper methods for determining cause an d origin. Negative corpus evidence is specifically warned against in section 17-4.
In the absence of physical evidence of an incendiary fire, the investigator is strongly cautioned against using the discovery or presence of other evidentiary factors, in developing the hypothesis, forming opinions, or drawing conclusions concerning th e cause of the fire.
The "other evidentiary factors" referred to in 17-4 refers to materials developed in the pursuit of suspect development and identification such as a motive related to financial stress, over insurance, an attempt to hide evidence of another crime, a pat tern of earlier fires, timed opportunity and the like. None of these secondary factors are present in this case and none of them would have been further developed in a well-run case in which there was no physical evidence of incendiary cause. It has bee n remarked that Ms. Hinson is a poor woman but her relative position in terms of ability to provide sustenance to her family was better at the time of the fire than it had been previously. The fact that Ms. Hinson had previously allowed the adoption of t hree children in a period of financial difficulty is proof enough that she would have been easily able to find her child a suitable home had she wished to free herself from responsibility.
The physical evidence referred to in 17-4 as a prerequisite for proceeding with other evidentiary factors includes a specific list of items divided into the categories of "incendiary fire indicators" (17-2 et seq.) and "potential factors not related t o combustion (17-3 et seq.).
Incendiary fire indicators. The list of incendiary fire indicators in NFPA 921 includes: incendiary devices, multiple fire origins, trailers, lack of expected fuel load or ignition sources, unusual fuel load or configuration, inconsistent burn injuries, delay devices, presence of ignitable liquids, and excessive rate of fire growth. None of these indicators was present at the scene. Although it has been alleged that the arced, broken and melted wire at the heart of the fire was "merely a vict im of the fire," this subjective comment does not constitute a "lack of ... ignition sources" as is discussed later in this report.
Potential factors not related to combustion. The list of non-combustion factors includes: remote location, fires near service equipment or appliances, removal or replacement of goods, absence of personal items, entry blocked or obstructed, sabo tage to structure or fire protection, damage to fire-resistive assemblies, damage to fire protection systems, and open windows and exterior doors.
None of the elements of the three categories of evidence recognized by the NFPA as indicative of arson is present in this case. The subjective opinion that the damaged electric wire found at the scene was not the cause of the fire is dealt with in det ail later in this report. In the absence of dispositive indicators the presumption of arson rests solely on negative corpus reasoning. In some jurisdictions, the courts have specifically determined that negative corpus evidence is not an adequate basis for arson indictment, for example:
Commonwealth v. Moyer, 419 A.2d 717 (Pa. Super. 1980) ("Unsubstantiated suspicion that fire was caused by other than accidental means is insufficient to establish corpus delecti of arson; thus case should not have been submitted to jury)It is a common occurrence for fire investigators to misinterpret fire patterns in the absence of better evidence of arson such as the presence of an accelerant or incendiary device, multiple origins or a plausible motive, etc.. Frequently the smoke an d char marks lead to the mistaken conclusion that an accelerant was used, as recently happened in the Germaine Smith case. (State of North Carolina v. Germaine Smith, Wake County Superior Court, October 1996) In that case it was the erroneous interpretat ion of fire patterns which led to the mistaken assumption that the potential accelerant found near the scene had actually been used to initiate the fire. The Smith case has many similarities with that of Hinson in that subjective evaluation of fire patte rns led the investigators to ascribe an imaginary accelerant scenario. At least those investigators believed, albeit falsely, that they had evidence to corroborate their causation theory in the form of a near empty gasoline can within the house.
Smith, who never should have been tried, was acquitted. Attached to this report is a copy of an investigative report describing the errors made in prosecuting the Smith case using scientifically invalid data. This case study should be read by both th e defense attorneys and the prosecution because it describes a recent North Carolina trial in which many of the junk science errors leading to the indictment correspond closely to similar errors made in the early stages of the Hinson investigation.
Junk Science. The term "junk science" is used in this report in the same sense that it is now associated with the law being adopted in many states following the United States Supreme Court decision in Daubert. (Daubert v. Merrell Dow Pharmaceut icals, Inc. 509 U.S. 579 (1993)) The language of the new rules stresses the role of the courts in acting as gatekeepers to prevent the submission to juries of evidence based on alleged scientific methods which are not generally accepted in the relevant sc ientific community or backed by credible research. A brief treatise on the clarifying language of Daubert with respect to the older and widely accepted Frye laws is attached.
In the Hinson case the fire scene showed only two significant areas of burning, the closet in the deceased child's room and the attic. The area of heaviest burning was in the attic, where sections of rafters and joists were completely consumed immedia tely above the closet. These wooden beams were the most massive specimens destroyed in the fire and therefore undisputedly identify the area of deepest penetration and charring. Both the closet ceiling and the roof, each comprised of less substantial to ngue-in-groove wooden boards, were penetrated creating comparably sized, approximately vertically aligned holes. Clearly, fall-down, the dropping of flaming and embrous debris, would have occurred in the closet from the disintegrating ceiling and roof ma terials during the fire regardless of the point of origin. In addition to the logical assumption of the existence of fall-down there is direct evidence of its presence in the form of a broken electric wire the ends of which are bent into the closet ceili ng hole from the force of the falling debris.
THE OFFICIAL INVESTIGATION
Cursory official investigation. The official scene investigation occurred on October 20, 1996, and was conducted by SBI Special Agent Matthew White and ATF Special Agent Frank Malter. The two officials filed their report with a finding of ars on only four days later on October 24. Despite the heavy damage to the roof they remarked that the "interior of the closet exhibited the greatest fire damage." Apparently, they included the ceiling joists located in the attic as part of the closet interi or. Noting that the underside of the closet shelf and the right wall were respectively more heavily burned than the top of the shelf and the left wall, they concluded that the origin was beneath shelf in the right portion of the closet. In order to elimi nate any possible electrical origin they inspected the electrical wall outlet in the closet and declared it free from internal heating and therefore not the cause of the fire. Based on this limited set of observations the official investigation team made the negative corpus conclusion that "All accidental causes in the area of origin were eliminated. This was an incendiary fire ..."
Electrical failure indicia overlooked. If agents White and Malter had looked more closely for electrical causes they might have noted the broken wire of the Attic Romex cable bent into the closet hole. This wire, as was later reported by insur ance investigators, had three independent areas of damage which could constitute signatures of an electrical origin:
1) The break in the wire could have existed prior to the fire and caused a prolonged glowing contact yielding up to approximately 25 watts of local energy release.The careful inspection of the electrical outlet in the closet shows that the agents were aware of the necessity of eliminating electrical origins to support any effort to lend credibility to a negative corpus conclusion, but apparently overlooked the m ore important Romex cable located at the center of the fire.
Possible weather influences. October 20, the morning of the fire marked the first cold snap of the year as can be deduced by the first usage of two electrical heaters in the subject house. The effects of the sudden drop in temperature after m ild weather raise two questions that should have been addressed by the investigators. The first obvious question is whether there was a relationship between the increased load on the electrical circuit and the second whether the onslaught of cold weather caused condensation leading to increased propensity to arc tracking.
Ambiguous fire patterns. The observation by Agents White and Malter of heavier burning to the underside of the closet shelf and right closet wall is not proof that the fire started in the lower right portion of the closet because it does not ta ke into account the insulating properties of the materials on the shelf and the inevitable effects of fall-down on the particular fuel load under the shelf. The shelf had upon it in left-to-right abutting order the following items: folded bedding (Pillow s, spreads, blankets, comforters), stacks of towels and washcloths, an empty popcorn tin, and a box of six glasses. Folded, stacked fabric items are poorly combustible and their underlayers act as excellent insulators. The glasses and tin box are also g reat fire barriers. Below the shelf there were hanging clothes, which, because of their loose vertical orientation act as accelerants, burning rapidly as long as oxygen is available. Any fall-down from the ceiling or roof would ignite the hanging clothi ng and produce an intense fire under the shelf. The heavier, rapid burning would occur under the portion of shelving that was immediately above the garments and to the wall which was closest to them.
The ventilation problem. It should be noted that the total oxygen available in the original closet with the doors closed and at room temperature would be only about 2 pounds and this amount would be reduced rapidly by the thermal expansion of t he heated gases. It is unlikely that the available oxygen could support the flaming combustion of more than about a pound of clothing before it was limited by the low but unknown ventilation by hypothetical imperfect joints in the construction. The wood would not become significantly involved in the fire until the free oxygen was depleted. The fallen clothing would continue to smolder and compete with the wood for any inwardly diffusing oxygen. This condition would produce an extremely smoky atmospher e and slow combustion, resulting in a smokehouse effect with the smoke horizon reaching the floor. In this case, the horizon is well above the floor and thus not consistent with an origin in a closed closet.
If the closet door was open even slightly and the fire originated in that closet, the combustion of the hanging clothing would be extremely rapid and the wood would become quickly involved. The open slit of the door would serve as an intake port below the neutral plane and as an egress vent above it, rapidly involving the wooden exterior wall and wooden ceiling immediately above the closet door followed by the rest of the ceiling and upper wooden walls of the bedroom. The neutral plane of a door open ing is the horizontal level at which the pressure in the closet is equal to the pressure outside and at which no net gas flow occurs. Above that level, buoyant, hot gases and flames flow out and are replaced by cold air flowing in below it. The ventilat ion factor which controls the rate of flow is given by the product of the opening area and the square root of the height Since the agents noted only an area of heavier ceiling damage about 3 feet from the doors, it is obvious that there was no prolonged b urning in that room and that, regardless of the fire origin, the closet door was not open but closed. The report of the very fast fire occurring in the bedroom proper but producing only moderate damage is consistent with the fire having involved the room only during the short period when it broke through above the closed closet door, was observed by Hinson and very shortly thereafter extinguished by the fire department.
Burn patterns consistent with an attic fire. It is the contrast between the attic and room damage levels that shows the fire burned long before creating an open path to the room. As a starting rule the area of greatest fire destruction is pres umed to be coincident with the area of origin of a fire. In this case the damage is progressively less in going from the attic to the closet to the bedroom. Obviously, ventilation factors also play a role in degree of fire damage but there is no viable way to estimate these effects inside the closet within the framework of the arson proponents theory in which the investigators could not even agree on whether the closet door was open or shut. At the very least, the fire patterns are completely consiste nt with an origin in the attic as supported by the monotonically declining damage levels progressing from attic to closet to bedroom to other upstairs portions to downstairs.
White and Malter observed a circular burn pattern on the bedroom floor "consistent with the right closet door," implying that the door was open in accordance with Mangini's finding on the same day. Circular patterns are common phenomena at fire scenes and arise from various sources discussed in NFPA 921 4-17.7. All it takes is a bit of flaming debris from the ceiling or a light fixture to start a spot fire which will expand in circular form. Noting the delamination of ceiling materials in the adjacen t, less damaged bathroom, it is obvious that there was probably similar material dropping in flames in the bedroom. The light fixture in the bedroom was also missing after the fire.
The case for an origin in the closet does not comport well with the burn patterns at the scene but a scenario beginning with an attic fire followed by any amount of fall-down into the closet is well supported by the evidence. The presence of burning f all-down is a given any time wood is burned away as is observed commonly under fireplace grates. The closet ceiling was no exception. While there is always room for argument in any analysis of fire patterns, it is quite clear that there is no defensible support for a negative corpus finding of incendiary origin by any process of elimination which fails to take into account the reasonable possibilities of an attic fire induced by the mechanisms previously discussed.
THE INSURANCE COMPANY INVESTIGATION
Perfunctory dismissal of wire arc ignition. On the day of the fire, the official investigation was immediately followed by a site survey by a Nemex investigator, Mr. Frank Mangini. Nemex was at that time a subsidiary of the house insurer, USF &G. It is not uncommon for police and insurance investigators to mutually support each other in cases where one or the other believes to have found indications of arson. ATF agents and state fire officials are often called upon to testify in civil trial s on behalf of insurers, and insurance investigators frequently aid the authorities in prosecuting alleged arson. In this case, Mr. Mangini quickly concurred with the prior investigation by Malter and White, also using negative corpus reasoning to arrive at the conclusion of incendiary origin. His conclusion was tersely expressed as follows:
"With the elimination of available ignition sources it must be concluded that this fire occurred due (sic) to some type of human action."
Mr. Mangini apparently looked at the scene in more detail with respect to electrical origin and noted the existence of the Romex cable. However, he failed to mention in his report that one conductor was broken and he dismissed the wire as irrelevant, apparently without scrutinizing it closely. His perfunctory examination led to the blunt and incorrect negative corpus conclusion expressed in a single sentence:
"This conductor showed no signs of electrical activity that would suggest it was the cause of the fire."
Discrepancy about the closet door. Five days after Mangini's inspection, USF&G brought in a second expert, a mechanical engineer, Michael Sutton. Mr. Sutton essentially verified the opinions of Mr. Mangini with respect to fire travel except f or the position of the closet door. Mangini, who investigated the scene on the same day as the SBI and BATF agents, had determined that it was open whereas Mr. Sutton opined that it was closed, apparently unaware of the contrary finding because Mangini's report had not yet been written. Sutton also noted the upward burn patterns on the walls of the closet (consistent with fall-down) and added a comment that the ceiling joists were burned "from the bottom to the top."
Misinterpretation of fires which burn down and up. The upward burn marks of the ceiling joists were clearly mentioned in the context of the fire having originated in the closet. Every fire investigator is taught to identify the direction of t ravel of a fire in terms of the geometry of burned areas. A fire burning down through a horizontal hole tends to produce scalloped edges and joist ends contoured such that the wider portion faces the progress of the fire. Thus the shape of the hole is a n indicator of whether the fire burned in an upward or downward direction. However, like most fire pattern evidence it is not a simple or straightforward as it appears at first glance.
Effect of fall down on burn patterns. Any fire starting in the attic would in its natural course produce fall-down after it breached the ceiling. The heavy and available fuel load in the closet would then burn much more rapidly back up through the same hole, enlarging the diameter thereof and destroying the original downward indicia. This principle is well accepted in fire science and is articulated clearly in NFPA 921 4-16.4.2 which warns about drop down (fall down) leading to false points o f origin and NFPA 921 4-17.7.1 which specifically warns about the reversal of fire through horizontal holes which creates inverted fire patterns. The overall principle expressed in the two sections is that fire patterns may reflect only the last directio n of travel by the fire. Neither of these two well-accepted and extremely important NFPA admonitions was taken into account by either the official investigators or the insurance investigators.
Three areas of wire damage. It appears that the primary motivation for USF&G in bringing in a second investigator was to reinforce the unsupported comment by Mr. Mangini that the wire showed "no signs of electrical activity that could have caus ed the fire." Mr. Sutton noted damage which was unreported by Mr. Mangini in the form of an arc-marked section of the wire, a separate portion which had melted and a break in one conductor. The arc, melt and rupture were in three distinct sections. The break was described as "mechanical" in nature. Like Mr. Mangini, Mr. Sutton peremptorily dismissed the arced wire as a victim of the fire.
The "Arc, cause or effect" problem in science. The frequent question of whether an arc occurred before or during a fire is often heard expressed as, "Arcs cause fire and fires cause arcs." In many cases an investigator will make the claim that he can tell the difference by simply looking at an arc bead and noting the contact angle at the interface of the bead (if it is a bead as opposed to some other form of arc damage). The fact is that there is no accepted method of proving that an arc was the result of a fire. The converse situation in which an arc starts a fire can sometimes be shown by microscopic examination of the grain structure of the portion of the wire adjacent to the bead to verify that the orientation induced in the original dr awing process has not been destroyed. The potentially exculpatory evidence of cold arcing only be verified by laboratory analysis. If simple field observation could discern the history of a fire-treated wire, there would not presently exist so much resea rch into potential sophisticated laboratory techniques for accomplishing the same goal.
The state of the art in arc causation. The leading research in the area of arcs and other electrothermal fire-igniting phenomena appears to be that presently being conducted in Japan. The international state of the art in arc-related phenomena in relation to fire cause can be best judged by reference to that body of work summarized in the document "Japanese Reports on Electrical Fire Causes" in which various segments of the effort are summarized and translated by Yasuaki Hagimoto (National Re search Institute of Police Science, Japan) in connection with his visit to VTT, the Technical Research Center of Finland, in February and march 1996. This very long and complex Japanese summary details numerous studies under the headings "Discrimination between Primary and Secondary Arc Marks on Electric Wires," or "Research on First and Second Fused Mark Discrimination of Electric Wires." The combined research summaries show a recent and ongoing scientific effort to find a means of determining if an ar c signature on a wire is primary, i.e., occurred before a fire, or secondary, i.e., occurred during a fire as a result of a direct short circuit or of arc tracking.
Potential electrical causation not considered. Additional sections of the extensive Japanese report on electrical causation contain detailed descriptions of experimental work on low energy glow phenomena which occur when the ends of a broken s ingle conductor are in contact as may have happened in the Hinson fire. Neither the arc nor the glow results have yet been incorporated in standard references for fire investigation. The mere existence of this intense research underscores the lack of an y scientific validity in the claim of the insurance investigators that they can eliminate the possibility that the observed arc mark, broken wire or melted section of the subject wire was responsible for the fire. The perfunctory observation of electrica l signatures, melting or breaks in a wire in a routine fire investigation cannot do what the most sophisticated laboratories in the world have not yet achieved.
In order to learn the details of the physical evidence in this case, it was necessary to study every available document from every source. The work necessarily included the reading and comparison of the statement and interviews of various fact and cha racter witnesses. I will not comment on the materials apparently gathered in the unsuccessful search for some motive for a mother to murder her youngest child. Like any other case, this one includes the usual items of gossip about the personal idiosyncr asies of the defendant. I would like to mention one example of misleading information which might unduly influence other readers if they fail to study all of the interviews completely and relate them to one another.
* * * *
Unexplored communication errors by police and fire departments. In the interviews with emergency personnel present at the fire scene there are misleading instances of allegations that the defendant failed to inform the authorities of the locati on of the children or, worse, deliberately misled them as to the location of the child who died. However, after careful analysis of all the interviews it becomes apparent that Ms. Hinson did, in fact, accurately disclose the pertinent information to nume rous participants at the scene but that the information was not properly further communicated between the agencies and individuals who needed the knowledge to conduct the rescue operation efficiently..
Conflict of interest among witnesses. The small size of the Tabor City community dictates that there exist personal relationships between the members of the police, social services, fire and emergency team members and probable frequent biologi cal and marital connections. Errors in communication among three of these groups at the fire scene may have unintentionally hindered the rescue of Joshua Cade before he died of carbon monoxide poisoning. The aftermath of an unfortunate situation such as this can put pressure on the individual participants to protect their colleagues from embarrassment by unconsciously focusing attention elsewhere.
Clearly, there is no physical evidence in the Terri Hinson case that will support the conclusions of the prosecution experts with respect to an incendiary origin of the fire. The methods used by the investigators in reaching their negative corpus arso n determination do not withstand elementary scientific scrutiny based on widely accepted principles of fire investigation. The interpretation of the burn patterns by the experts for the prosecution and the insurance company ignores even the most fundamen tal rules promulgated in the most authoritative and generally acknowledged fire investigation reference work, NFPA 921. The failure to recognize that a fire burning downward into the wooden ceiling would be followed by an upward fire, which would eradica te the original downward pattern and replace it with an upward pattern, is an egregious oversight.
The arbitrary dismissal of the now missing, broken, melted and arced wire from the heart of the fire over the closet as being simply a fire artifact is based on an erroneous belief that such a determination can be made with any degree of certainty by m ere visual observation. If the wire were available it would be possible to submit it to a laboratory to determine if it shows indicia of abnormal electrical activity prior to its involvement in the fire. Although the fire was declared to be arson by st ate investigators and officially reported as such on October 24, 1996, no attempt was made to secure the fire scene or otherwise preserve potentially exculpatory evidence until 5 weeks later on November 26, 1966, when the premises were belatedly locked. In addition to the glaring problem of the missing wire there is the more general problem of the failure of early investigators to record their activities prior to the poorly documented investigations of the state, federal and insurance investigators. Pic tures of the area of origin show extensive alteration of the surroundings without any attempt to catalog the evidence. The normal procedure in an arson investigation would include a careful sifting of the debris and an equally careful listing of the loca tion and characteristics of the materials present. The needless destruction of potentially exculpatory evidence in a case which includes early allegations by the state of arson is simply unacceptable.