Allstate Insurance v. Sunbeam Corp.

PLUNKETT, District Judge. Because the Magistrate’s Report took the form of a dismissal rather than issuing sanctions in the form of barring testimony, we review the record de novo and we adopt Magistrate Judge Bobriek’s Report and Recommendation dated June 23, 1994 and order that Sunbeam’s motion of sanctions for evidence spoliation is granted and this complaint is hereby dismissed with prejudice and with all costs to defendant. REPORT AND RECOMMENDATION OF MAGISTRATE JUDGE EDWARD A. BOBRICK, United States Magistrate Judge. Before the court is a Motion for Sanctions for Evidence Spoilation [sic] 1 filed by defendants SUNBEAM CORPORATION and SUNBEAM LEISURE PRODUCTS COMPANY, a division of SUNBEAM CORPORATION’S (“Sunbeam”). I. FACTS On July 4, 1990, a fire damaged the house of Benito and Eutimia Garcia. The fire has precipitated two suits. The Garcias filed a products-liability suit No. 92 C 4897, in this court, against defendant Sunbeam Corporation (Sunbeam) alleging that a propanefueled gas outdoor grill manufactured by Sunbeam was defective or unreasonably dangerous and responsible for the fire. The Garcias’ insurer, plaintiff Alstate Insurance Co. (Alstate), *1269 subsequently exercised its right of subrogation under the Garcias’ homeowner’s insurance policy and brought this companion suit against Sunbeam. Both suits were initially filed in the Circuit Court of Cook County, Illinois, and were removed by Sunbeam pursuant to 28 U.S.C. § 1441(a) on the basis of diversity of citizenship under 28 U.S.C. § 1332. The suits have not been consolidated. Sunbeam has moved to dismiss this action because Allstate, in its investigation of the fire, failed to preserve certain evidence Sunbeam contends is essential for its defense. The motion was referred to this court for hearing and for a Report and Recommendation. A hearing was held November 2 and 3, 1993, at which Sunbeam’s expert, William Baynes, testified; the testimony of other witnesses was offered through depositions. A home video tape of the fire scene taken the same day of the fire and an Allstate investigator’s photographs of the scene were also received into evidence. The record in this matter was closed on February 2, 1994, and the matter fully briefed on March 18, 1994. After considering the testimony and examining the video and photographic exhibits, we recommend that the motion be granted. II. COMPETING THEORIES OF THE CASE The testimony and its significance will be easier to understand if we briefly state the parties’ theories. The parties agree that the fire that damaged the Garcias’ house was caused by a release of propane gas at the grill, which was then ignited by the grill, creating huge flames erupting against the back wall of the house, entering the basement and kitchen windows, and shooting into parts of the Garcia’s backyard. The grill itself burns propane, a form of LP (liquefied petroleum) gas, which is stored in a cylinder (tank) intended to hold twenty pounds of liquified gas under pressure. Allstate contends that the fire started as a leak somewhere in the “gas train,” the system of pipes, valves and hoses conveying gas from the cylinder to the burners in the grill. The fire grew until it overheated the propane tank, causing it to vent a large quantity of propane. Allstate maintains that had the grill been equipped with a thermal shutoff or safety disconnect device, the fire would not have overheated the tank and the catastrophic fire would have been prevented. Sunbeam, on the other hand, contends that it is more likely that a spare tank of propane had been stored beneath or directly behind the grill, and that this spare tank had been overfilled with liquid propane, leaving insufficient room for expansion of the liquid propane, said expansion being brought on by heat from the sun, the hot ambient air, and the nearby operating grill. This expansion of the liquid, in turn, filled the tank beyond its capacity causing the relief valve to open, releasing the liquid propane as a gas into the underside of the grill, where, by the natural upward flow of air (convection currents), it was drawn in contact with the burners. The gas would have ignited from the then burning burners in the grill, generating even more heat onto the second tank, thereby increasing the pressure within that tank until there was a massive release of flaming gas from its relief valve. Heat from the flames coming from the relief valve of the second tank would then have caused the service tank connected to the grill to overheat, become overpressurized, and to vent and flame as well, all of which caused the inferno that ran from the swimming pool to the back of the Garcias’ house. In support of this theory, Sunbeam points to photographs of the scene taken by investigators two days after the fire, as well as a home video tape taken by the Garcias approximately 8 hours after the fire, all of which show a second propane tank among the remains of the grill, in addition to the service tank used for fueling the grill at the time. In addition, the photographs showed possible burn marks on the second cylinder as well as burn patterns on the house and backyard swimming pool that suggests that two burning propane tanks were involved. Sunbeam argues that if its theory of what happened is correct, then it could not be at fault because (a) it was not responsible for the second tank being overfilled and being stored in close proximity to the grill, contrary to the user instructions, and (b) the *1270 safety devices suggested by Allstate would not have prevented the catastrophic fire since they would not have interfered with the normal operation of the relief valve which is designed to vent in the event of overpressuri-zation (as caused by the significant heating in this case). Sunbeam has a major difficulty, however, in proving its theory: first, the second tank shown in the photographs, and clearly depicted on the video tape, cannot be found or examined. Second, the grill frame and wood accessories, which could have shown burn marks and patterns supporting Sunbeams theory, are also nowhere to be found. There is no secret, however, behind the absence of this evidence. Shortly after the fire, an Allstate investigator, in the performance of a fire investigation, preserved only the service tank that was connected to the grill, the connecting fittings, the remains of the regulator, and the remains of the burners (but not the separate third burner above the tank). He directed that everything else be thrown away, which included — the third burner, the grill frame (the firebox had melted), any wood accessory remnants, and the second propane tank shown in the video tape and photographs. Sunbeam contends that the absence of this physical evidence cripples its defense. Since the evidence was discarded at Allstate’s direction, Sunbeam asks the court to sanction Allstate by dismissing this suit. Allstate responds that the lost evidence was not material because there was, in fact, no second tank near the grill. Allstate supports its denial of the existence of a second tank near the grill with eyewitness deposition testimony to the effect that only one tank was involved in the fire, and that the second tank on the premises was located a considerable distance away from the grill and was, in any event, empty. According to Allstate, in the confusion following the fire, somebody must have moved the second tank near the remains of the grill. On the day of the hearing Allstate offered a motion in limine seeking to bar testimony concerning the possibility that a second tank caused the fire, contending that Illinois law prohibits the use of expert testimony to “reconstruct” an accident according to a hypothesis denied by all of the eyewitnesses. The motion was taken under advisement, and we will address it below. III. TESTIMONY CONCERNING THE FIRE With these theories in mind, we review the evidence. On the day of the fire, July 4, 1990, the Garcias’ daughter and son-in-law, Lourdes and Bellarmino Sagols, were visiting them. Lourdes was sunbathing in a lounge chair in the backyard, while her son played in an above-ground swimming pool. The grill was located on a concrete patio and stood three or four feet from the back (east) wall of the house. To the north was a gangway between the Garcias’ house and the neighboring house, and the back door of the house opened onto this walkway. The grill had a large firebox and a hood that covered it. The gas burners in the firebox were covered by a layer of ceramic briquettes. There were control knobs in front of and underneath the firebox. On the left side of the grill was a propane cylinder mounted on a bracket and above it an additional gas burner similar to that on a kitchen range. The frame of the grill had an attached wooden front panel and was open on the other three sides. (See photograph of advertising material, PI. Ex. B). At approximately 11 a.m., Bellarmino Sa-gols lit the grill. After satisfying himself that the grill was lit and working properly, he left to go to the store. About ten minutes later, Lourdes happened to see flames coming from the grill, and went over to see what was wrong. Her deposition testimony is equivocal as to where the flames were coming from. This is partly because the deposition transcript indicates that she pointed to photographs, with no record of where she was pointing, but it also appears she was uncertain. At first, she indicated that the flames came from below the tank: “They were coming from underneath over here (indicating). From the back part underneath the tank.” (Lourdes Dep. at 21). Later, after a question from the Garcias’ counsel, she said they came from the top of the tank: *1271 A: [T]he flames were splattered all over. They were coming this way, this way, from the top, back, everywhere. Q [Allstate’s counsel]: You could see the flames coming up from underneath the grill? A: Yes. They were coming from everywhere. Q [Garcias’ counsel]: Was it from the grill or from the tank? A: It was from the tank. [Allstate’s counsel]: Did you actually see exactly where on the tank? A: When I saw the first fire coming up, I wasn’t sure where it was coming from. So I went over to the side, and that’s when I saw it was coming from the top part of the tank. Whatever that thing is, I’m not sure. [Allstate’s counsel identifies and shows witness photograph] A: It was coming from over here, somewhere in here (indicating). Q: So you’re just saying generally— A: The top area. Q: The top valve area? A: Yeah. Somewhere around there. That’s where the fire was. Q: And this cylinder or the tank where you saw the fire coming from, that was on the left side of the grill as you faced it? A: Yes. Q: Did you see the fire coming from anywhere else underneath the grill? A: It was just coming from here (indicating). Q: There wasn’t — you couldn’t see it anywhere else. Do you know if the grill was still lit inside when you saw the fire underneath it? A: What do you mean? Q: Inside the grill itself. Did you see any fire inside the grill? A: All I know, everything was just coming from the bottom all the way through the top and everything happened so fast, in a couple of seconds, (Lourdes Dep. at 24-26). As Lourdes stood in front of the grill, a neighbor, Leonard Allen, who had seen the flames from his backyard, ran up to her. He told her to turn the grill off, but she saw the control knobs had already melted. She started screaming. Allen told her to get back. As she turned, she heard a loud “shoosh” and flames shot out. Her hair was burned. She crawled underneath the shooting flames in order to get out of the yard, and went into the house. She did not see Allen again. (Lourdes Dep. at 22-28). Allen’s testimony was not offered. When asked if there was more than one tank under the grill, Lourdes first said she had no idea, then said she thought there was only one. She said there was another tank they used to keep fifteen or twenty feet away near the fence. (Id. at 29-30). Another neighbor, Ronald Mancini, ran over. He apparently arrived after Lourdes had escaped the flames, as he saw nobody in the yard. He testified he had heard a loud “whoosh,” turned around and saw flames as high as the eaves of the house. There was a large jet of flame shooting straight up from the tank valve, and a smaller jet pointing south. (Mancini Dep. at 9, 16, 22). Mancini was positive that there was only one tank under the grill. He said he was sure because, as a fireman, he was trained to lóok for flammable materials near a fire. (Id. at 9-10). Mancini ran into the house and escorted Mrs. Garcia out, then went back in and saved the dog. Returning to the back yard, he told a neighbor to stop spraying the grill with water, because water would only spread the flames. Mancini did not remember the appearance of the grill, other than seeing it engulfed in flames, nor could he remember whether the burning propane cylinder was on the left or the right side of the grill. (Id. at 49-51). He confirmed that he only saw one gas cylinder, and the flames came from that cylinder. (Id. at 51). He speculated that the second tank probably was put together with the other debris by firefighters, although he admitted that it was not standard procedure to *1272 put a canister of fuel in fire debris. (Id. at 54). Bellarmino could not recall whether there were two propane tanks in the backyard, but remembered that there was not a second cylinder stored underneath the grill. (Bel-larmino Dep. 40-42). According to Benito Garcia, there was an empty propane tank about fifteen feet from the grill on the south side of the yard, located where the garbage cans were stored. (Benito Garcia Dep. at 102, 106-107, 120-21). He testified that the photographs showed the grill where it had been prior to the fire, but the empty tank had been moved next to the remains of the grill. He did not know who moved the other tank, and he could not identify which tank was the operating tank and which was the one that had been moved. (Id. at 107-110). A home videotape recording made the evening of the fire, showing the fire damage, was placed in evidence. 2 In this tape, water can be seen dripping from the ceiling and there appears to be smoke in the basement, confirming that it was taken within a few hours after the fire was extinguished. The shots of the fire scene are blurred and include no closeups. Nevertheless, a second tank located next to the front of the gas grill, where the photographs show it to be two days later, is clearly visible in the video tape. (Tr. 146-149; Def.Exs. 13, 14). • On July 6, 1990, two days after the accident, Edward Janak, Allstate’s adjuster, came to look at the scene of the fire. He was advised that the site had not been disturbed. (Janak Dep. at 38-34). Allstate’s contract photographer arrived a short time later, and took a series of photographs that were admitted in evidence. 3 (PI. Exs. A-JJ, 2, 5,12, 13). Some of the photographs show the fire scene with the windows boarded shut. (The video shows them uncovered with the glass missing). (Del. Ex. 13, 14). The photographs show soot deposits on the swimming pool and the ladder, as well as on the back wall of the house. The photographs clearly show two gas cylinders, one standing to the left of the grill frame and the other lying on its side behind and to the right of the grill frame. (See PI. Exs. 2, 5, 12, 13, W, B). Two of the photographs, namely Plaintiffs Exhibits 12 and 13, best depict the two tanks near the grill frame. In these photographs, the tank shown on the left of the photographs is the service tank, whereas the tank shown on the right is the second tank. Although several close-up photographs were taken of the service tank, none were taken of the second cylinder. Sunbeam maintains it is this second cylinder that probably vented propane gas into the burning fire box of the grill, causing the ensuing conflagration. Significantly, the photographs do clearly show charring on the top and on the shoulder of the second tank. The photographs also show the second tank to have discoloration or soot-ing on the left side of the shoulder, and a discolored portion on top. (PL Exs. 2, 12). When Janak (Allstate’s investigator) learned that a gas grill was involved in the fire, he asked for a mechanical engineer to come to the scene. The engineer, Todd La-vieri, examined the scene of the fire and determined that the fire had indeed been caused by the grill, rather than something inside the house. He took with him the propane cylinder that had been in use, the remains of the regulator, one control valve and the burner, leaving the rest of the remains of the grill at the scene. (Lavieri Dep. at 69). After examining these parts, as well as a similar intact grill, Lavieri concluded that the fire had probably been started by a gas leak, but he was unable to determine where that gas leak had been. He could not conclude that the grill had been defective. (Id. at 73-76). Allstate arranged for the *1273 disposal of the rest of the remains of the grill, including the frame, the other grill debris caused by the fire, and the second tank as well.- (Janak Dep. at 83-84; Garcia Dep. at 148). IV. ALLSTATE’S MOTION IN LIMINE At the outset, we note that the hearing held on Sunbeam’s motion did not deal in any way with resolution of the disputed material facts as to the cause or responsibility for the fire. The instant hearing dealt solely with the issue of defendant’s ability to mount a defense in the absence of certain destroyed evidence, i.e., grill frame, second propane tank, etc. — key evidence as- proclaimed by defendants. Accordingly, the expert testimony introduced by the defendant, and considered by the court, was not used in any manner to reconstruct the facts of the case as those facts might have been perceived and described by certain eye witnesses. Instead, the expert testimony was utilized to ascertain the nature of the diminution of defendants’ ability to forward and maintain a cogent defensive theory regarding the cause and origin of the fire, in the absence of such evidence. We find plaintiffs argument in objecting to defendant’s expert erroneous and misplaced, since the expert’s testimony was not used to reconstruct any fact vis-a-vis the testimony of the eye witness. Indeed, we have not used, nor do we intend to use, for the purpose of these limited proceedings, the expert’s testimony to impugn eye-witness testimony, credible or otherwise. We rely on the expert’s testimony solely- to assist us to understand the import and ramification of the absence of the second propane tank, the frame of the grill, and the wood rack remnants, as that absence would affect defendant’s ability to mount a defense to the allegations set forth in the complaint. Accordingly, we easily conclude that defendant’s expert and opinion testimony is well within the framework of Federal Rules of Evidence 703, and plaintiffs objection to such testimony is, of course, on this basis, overruled. For the sake of completeness, however, we will review plaintiffs objections. Allstate, in its motion, seeks to exclude Sunbeam’s expert testimony that the second propane tank shown in the photographs could have been the cause of the fire. It argues that such testimony would be inadmissible speculation because there is no testimony that the photographs reflect the scene before objects had been moved by firefighters and others. Nevertheless, if we were to consider plaintiffs objection within the context of “reconstruction of evidence,” the first thing we would do is take judicial notice of Newton’s first law of motion, that an object will remain in place “unless it is moved. Allstate has not offered a plausible reason — or any reason at all — why anyone would have taken a propane tank from its place near the south fence and put it in the middle of the fire remains of the grill. On the other hand, it is quite plausible that the Garcias stored a spare tank of propane beneath the grill. It is not unreasonable to conclude that this would be a logical place since its convenience of access might win out over caution, especially if one disregards or is unaware of the risk of escaping gas due to over pressurization. More importantly, though, we note that if the second propane tank had been near the fence at the time of the fire, as described by some of the witnesses, and someone moved it later, it would not account for the charring and soot on the top and shoulder of the tank. One could hypothesize that it was caused by the jet of flame that scorched the swimming pool and the swimming pool ladder. If this had been the case, and the tank had been near the fence where Benito Garcia said it was, we would expect to see some soot or fire damage on the chair and the garbage cans (which appear to have been made of plastic) that were in the same area. There was no such soot or fire damage to be found at the location of the garbage cans. (See PI. Ex.“ 12). Allstate emphasizes that no eyewitnesses saw the second tank near the grill at the time of the fire. Eyewitness testimony is not always reliable, however, particularly almost three years after the event. Bellarmino did not recall seeing a second tank before or after the fire. Lourdes’ testimony was equivocal as to the origin of the flames and she was in front of the grill where the front panel could have obstructed her view of a *1274 second cylinder beneath the grill had it been there. Benito Garcia testified unequivocally that the second tank was empty and was fifteen feet away from the grill, but he is an interested witness, and his testimony shows that he was aware of Sunbeam’s theory and anxious to disprove it. The strongest evidence against the presence of a second tank is Maneini’s testimony. Coming through the gangway he would have been able to see the grill from the side, so the front panel would not have hidden a second tank from him. He had not seen the grill before, however, and he may have believed that a spare tank stored under the grill was in fact the operating tank; he may have assumed that the flames were obscuring the connecting hose he thought was there. At his deposition he could not recall on which side of the grill the operating tank was located, despite the fact that he claimed he was trained to look for and notice such things. We conclude that photographs and video recording raise more than a reasonable possibility that the second tank was indeed located near the grill at the time of the fire, and the testimony is not so overwhelming as to defeat it. As earlier discussed, the question is not whether a jury verdict based on the evidence of the. post-fire photographs over the eyewitness testimony could be sustained, although we believe it could. The question is not what caused the fire, .but whether the lost evidence would have been material. An expert need not base his opinion on undisputed facts, only upon a version of the facts that has some evidentiary support. The photographs, as well as the video tape of the fire scene taken the same day of the fire, show the second propane tank adjacent to the base of the grill. This is enough of a foundation to permit Sunbeam to offer its hypothesis, through expert testimony, of a second over-pressurized tank located at the grill and explain how the lost evidence would have been material in proving or disproving it. Allstate’s second objection is that Illinois law does not permit Sunbeam to use expert testimony to reconstruct the sequence of events leading to the fire in contradiction to eyewitness testimony. Although we are not now addressing the merits, Allstate’s objection is germane to the question of materiality. If, at trial, Sunbeam would not be permitted to present its expert’s opinion that a second cylinder caused the fire, the unavailability of the second cylinder would probably not be material. It appears that Illinois law does not permit the use of expert reconstruction testimony when eyewitness testimony is available and the factual issues to be resolved do not require scientific knowledge beyond that of typical jurors. Peterson v. Lou Bachrodt Chevrolet Co., 76 Ill.2d 353, 29 Ill.Dec. 444, 446, 392 N.E.2d 1, 3 (1979); Palmer v. Craig, 246 Ill.App.3d 323, 186 Ill.Dec. 237, 239, 615 N.E.2d 1294, 1296 (1st Dist.1993). Opinions of Illinois Appellate Courts are in conflict in recognizing exceptions to the rule and whether exceptions are recognized at all. See Augenstein v. Pulley, 191 Ill.App.3d 664, 138 Ill.Dec. 724, 732-33, 547 N.E.2d 1345, 1353-54 (5th Dist.1989). Although an expert’s testimony would be helpful in establishing under what conditions a second, full, gas cylinder could have vented gas and caused a catastrophic fire, whether a second cylinder was present and where the flames were coming from are not matters requiring scientific expertise. This rule could bar expert testimony used to establish these facts. Sunbeam cites Robles v. Chicago Transit Authority, 235 Ill.App.3d 121, 176 Ill.Dec. 171, 182, 601 N.E.2d 869, 880 (1st Dist.1992), where the court held that it was not error to admit reconstruction testimony that contradicted eyewitness testimony. In Robles, however, the testimony of the eyewitnesses was equivocal as to whether the plaintiffs decedent jumped from a moving bus or whether the bus started to move as he exited. Here, none of the eyewitnesses saw a second cylinder near the grill. Nevertheless, there is unequivocal evidence of a charred second cylinder at the scene of the fire. Under these circumstances, we can only wonder what an Illinois court would do in permitting an expert to present a reconstruction based on the more than circumstantial evidence of the video tape and photographs that stand in contradiction to all eyewitness testimony. See Olson v. Bell Helmets, Inc., 195 *1275 Ill.App.3d 20, 141 Ill.Dec. 732, 734-35, 551 N.E.2d 1075, 1077-78 (1st Dist.1990) (reconstruction evidence may not be used to rebut the testimony of an eyewitness unless he or she is found incompetent or incredible, and reconstruction evidence may not be basis of finding of incredibility). We need not answer that question, however, since the rule argued by plaintiff is not binding upon a federal court exercising diversity jurisdiction in any event. Trans-Cold Express, Inc. v. Arrow Motor Transit, Inc., 440 F.2d 1216, 1221 (7th Cir.1971), cited by Allstate, merely affirmed the district court’s application of the Illinois rule as within its discretion. It did not, as we read it, adopt or endorse the Illinois rule. In any event, Trans-Cold Express was decided before the adoption of the 1972 amendments to the Federal Rules of Evidence when Rule 43(a) of the Federal Rules of Civil Procedure provided that admissibility of evidence in diversity suits would be governed by both state and federal law. (Id. at 1218, n. 2). The admissibility of evidence in federal court is governed by the Federal Rules of Evidence. If the state law rule is a rule of admissibility of expert evidence, it does not apply in federal court. If it amounts to a substantive rule of proof—e.g., that a plaintiff cannot recover on the basis of showing X plus Y without showing Z—it does. Stutzman v. CRST, Inc., 997 F.2d 291, 295 (7th Cir.1993); Barron v. Ford Motor Co. of Canada, Ltd., 965 F.2d 195, 198-200 (7th Cir.1992). It is clear that the rule barring expert reconstruction testimony when eyewitness evidence is available—whatever its limits and exceptions may be—is a rule of admissibility. It reflects Illinois courts’ mistrust of expert reconstruction, viewing it as an inferior substitute for eyewitness testimony. “Because of a preference for eyewitness testimony, prior holdings have placed restrictions on the admissibility of reconstruction testimony.” Peterson, 29 Ill.Dec. at 446, 392 N.E.2d at 3. Such a preference is a procedural rule a federal court is not obliged to apply, Stutzman, 997 F.2d at 295; according to at least one Seventh Circuit opinion, a federal court shouldn’t apply it, since all evidence, even eyewitness evidence, is ultimately inferential and probable. See Milam v. State Farm Mutual Automobile Ins. Co., 972 F.2d 166, 170-71 (7th Cir.1992). Under Rule 702 of the Federal Rules of Evidence, all we need to ask is whether the expert’s specialized knowledge will help the court understand the evidence or determine a fact in issue. Sunbeam’s motion requires that we address three issues. (1) Whether a second cylinder near the grill could have caused the fire; (2) if it did, whether the safety devices Allstate claims should have been employed would have prevented the damage, and (3) whether the missing evidence would have been probative in determining whether a second cylinder actually caused the fire. These are all appropriate matters for expert testimony, and we find the testimony of Sunbeam’s expert to be helpful in deciding them. We therefore deny Allstate’s motion in limine. V. SUNBEAM’S EXPERT TESTIMONY Sunbeam’s expert, William Baynes, is Director of Engineering for Sunbeam Outdoor Products. He has degrees in electrical and mechanical engineering and has been working with barbecue grills since 1976 and with LP gas equipment since the early 1950’s. He designs gas grills; he is familiar with the Sunbeam model grill involved in the fire, and he has investigated more than a hundred gas grill fires. We found him a knowledgeable and credible witness. (Tr. 256-257, 466). 4 Baynes considered the testimony of Lourdes Sagols and Ronald Mancini regarding flames coming from the grill, the melting of the control knobs, the loud “whoosh” and the eruption of flame against the house. Baynes testified that this would have been the result of a large discharge of gas and is not consistent with a small leak in the gas train. (Tr. 228-229). *1276 Baynes explained how such a discharge can occur. Propane tanks of the type used with the grill have a safety relief valve set to open at 375 pounds per square inch, which is about one and a half times the working pressure. At 120 degrees ambient temperature, the pressure in a properly filled tank will be about 240 pounds per square inch. At 150 to 160 degrees, the relief valve will start to release gas. As the pressure in the tank rises to between 320 and 340 pounds per square inch there is a gradual release called “weeping.” If the gas catches fire, it produces a lazy, licking flame. If the pressure in the tank continues to rise, at around 375 pounds per square inch the valve opens and there is a sudden, almost explosive release. (Tr. 261-63). Baynes conducted an experiment in which an overfilled tank was heated until it vented. At first there was a small flame confined within the collar at the top of the tank; when the relief valve let go, it suddenly projected flame like a flamethrower. (Tr. 288-89). Baynes .explained that a propane tank is supposed to be filled to about 80 percent of its volume. Proper filling can be verified either by weighing the tank when empty and then when full (the tank in question was intended to hold 20 pounds of liquefied propane), or by opening a valve at the top of a small tube that extends down into the body of the tank to the proper filling level. If that valve is opened while the tank is being filled, filling should be stopped at the point liquid, rather than gas, squirts from the valve. (Tr. 259-260). If a tank is overfilled, it can become liquid filled (expanded liquid fills all of the space in the tank) at a much lower temperature. Overfilled tanks have been known to vent while sitting in the sun or being carried in automobiles. Given enough overfilling, a tank could vent at the air temperature existing on the day of the fire at the Garcias’ backyard. The United States weather service showed the outdoor temperature to between 94° and 96° F. at the time of the fire (Tr. 265). Also, at the time of the fire, the sun would have been shining directly on the grill and the tanks. Clearly there was enough heat present the morning of July 4, 1990, to cause an overfilled tank to liquid fill and then to vent. If the spare tank had been stored close to the grill, convection currents would have drawn the vented propane gas through ventilation holes in the bottom of the firebox and into contact with existing flame, and this in turn would have propagated the fire back to the point of the leak. (Tr. 290-92). This additional fire at the spare tank would have created additional heat, increasing the temperature of the tank and forcing gas out even more violently. The heat from the fire could have also melted the plastic seal inside the control valve, permitting gas to discharge from the spare tank’s outlet as well. (Tr. 264r-65). Baynes explained that it was unlikely that the operating cylinder, (the service tank attached to the grill), had vented and caused the fire in this way. Even if it had been overfilled, the tank had been used previously, lowering the level of liquid propane in it. At the time the fire started the grill was in use, withdrawing gas from the tank and cooling the liquid propane in it by evaporation. This made it unlikely that venting from the relief valve of the operating tank started the fire, although if the fire started elsewhere it could have heated the operating tank and caused it to vent as well. (Tr. 266-67). Baynes testified that the evidence showed that two tanks were involved in the fire. He did not believe that the operating tank was the only one to vent, because the operating tank was held in a bracket on the grill frame in such a way that the relief valve was pointing left and gas would have been vented from that valve toward the swimming pool,' not toward the wall of the house. (Tr. 300-305). Baynes believed that both tanks vented gas. The operating cylinder vented toward the swimming pool, which showed visible soot and burn marks, while the second cylinder was responsible for the blast of flame against the house; burn marks on the building and the roof, as shown in the photographs, established this fact. (Tr. 330-335; PI. Exs. 5, 12). This hypothesis is not inconsistent with the witnesses’ hearing a single “whoosh” since the sound of the relief valve breaching on the operating tank would not necessarily *1277 have been heard over the sound of propane venting from the second tank. (Tr. 323). Baynes testified that Lourdes Sagols’ seeing flames beneath the grill that melted the control knobs before she heard the “whoosh” was consistent with his theory of the fire that there had been a spare second tank under the grill, overfilled with propane. As this overfilled second tank was warmed by the sun’s rays the ambient air, and the heat from the grill, there was then a slow, “weeping” leak of gas from the second tank’s relief valve that caught fire, causing the quiet flames she saw beneath the grill. These flames further heated the second tank until the relief valve fully let go. (Tr. 389). Baynes opined that if the fire had been caused in this manner, a quick disconnect device or a thermal shutoff valve, as proposed by Allstate, would not have prevented it, since it would not interfere with a release of gas from the relief valve. (Tr. 272). The deposition testimony of Allstate’s expert, Wendell Alan Bullerdiek, was consistent with Bayne’s testimony in all material respects. He testified that an overfilled cylinder can leak gas at moderate temperatures. According to a 1986 report, a 10% overfilled ■ cylinder can become “liquid full” at a temperature of 108 degrees, and with 5% overfill at 129 degrees. While he based his own analysis of the fire on witness statements to the effect that the spare tank was empty and located at a distance from the grill, he did agree that if a full tank had been stored under or near the grill, the fire could have occurred in the manner Baynes described — a “weeping” release of gas from the relief valve of the second tank that caught fire, heated the tank, and precipitated a violent release of gas that then caused the operating tank to vent as well. (Bullerdiek Dep. at 103-104, 130-33). Bullerdiek, in his deposition, disagreed with Baynes in that he believed that the pattern of soot deposition on the back wall of the house was consistent with a release of gas from the operating tank alone, but yet he agreed that either scenario was possible. He also agreed that if the fire had started in the manner, described by Baynes, a quick disconnect device or thermal cutoff would not have prevented the fire. (Id. at 145-46). Finally, Baynes testified if he were able to inspect the second tank he could have examined the control and relief valves to see if the tank did have propane in it, and whether it discharged in a firey blast. He would have been able to identify any burn marks on the tank. (Tr. 273-283). Baynes testified that had the frame and wood accessory remnants been available, burn pattern could have been discerned to better determine and establish the origin of the fire (Tr. 273-275). VI. PREJUDICE TO SUNBEAM Having established that Sunbeam’s theory of the cause of the fire is reasonably plausible, we ask whether Sunbeam has been prejudiced by the missing evidence. Baynes cogently testified that it is impossible to determine the cause or origin of the fire from the parts preserved by Allstate, (a view also held by Allstate’s fire investigator Lavieri) (Tr. 272). He testified that a reasonable mechanical engineer would have saved all parts of the grill as well as the second tank. (Tr. 273). The grill frame and the third burner would probably have had burn marks on paintéd surfaces that would be evidence of the source and direction of the flames causing them. Remains of wood parts and the back of the control panel would likewise have helped to determine the source of the flames. (Tr. 273-276, 281-83). If the second tank had been saved, an examination of the control valve and relief valve would show whether it had contained gas at the time of the fire, and had vented in the fire. It would also have permitted Baynes to confirm that a deposit on the top of the tank was solidified molten aluminum, indicating that the second tank had been stored beneath or adjacent to the grill all along (the firebox, which melted in the fire, was an aluminum casting). (Tr. 284-86). In Baynes’ opinion, the photographs were not an adequate substitute for the physical evidence. (Tr. 287). Allstate’s expert, Bullerdiek, agreed that had he been investigating the accident, he probably would have preserved the second tank, and he would have photographed and *1278 perhaps preserved the grill frame and control panel. (Bullerdiek Dep. at 134-37). We conclude that the missing evidence — the remains of the grill and the second tank — were highly relevant and material, and their absence significantly and irrevocably prejudices Sunbeam’s defense; this conclusion is inescapable. VII. SANCTIONS We now address the question of the appropriate consequences. Allstate does not dispute that Illinois law requires a plaintiff complaining of a defective product to preserve the allegedly defective product (or what remains of it), and that this is a substantive rule binding on a federal court in a diversity case to be decided under Illinois law. State Farm Fire & Casualty Co. v. Frigidaire Div. of General Motors Corp., 146 F.R.D. 160, 162 (N.D.Ill.1992) (Aspen, J.). In State Farm the plaintiff insurer investigated a fire at the home of its insured, and determined that the fire began in the insured’s dishwasher. The plaintiff’s investigator determined that the fire had begun at the right front corner of the dishwasher. After cutting out a part of portions of the wiring harness, a switch and switch mounting bracket from the right front corner, he told the insured she could dispose of the rest of the dishwasher, which she did. When the insurer filed suit two years later, it was impossible to ascertain from the retained parts alone what had caused the fire. The court granted the defendant’s motion barring the plaintiff from introducing any evidence as to the condition of the dishwasher, a result functionally equivalent to dismissal. Allstate argues that it was not obligated to preserve the second tank found with the remains of the grill, since it was not part of the allegedly defective product. According to Allstate, it was merely a piece of material from the accident scene, and Allstate no more had a duty to preserve it than to preserve, say, the swimming pool ladder. (PL Brief at 16-18). Allstate contends that it did save all of the parts of the grill that were material to its theory, which was the only theory supported by eyewitness testimony. (Id. at 18). The short answer is that Allstate did not preserve the entire product. It did not preserve the frame or the third burner, which might have helped determine from which direction flames had come. We believe Allstate had a pronounced duty to preserve the second tank as well. The rule requiring preservation of evidence rests on basic fairness, and there is no reason to limit it to the product itself. This is especially applicable to Allstate since it all along intended to proceed, in a subrogation suit. If a would-be plaintiff is in control of evidence he can reasonably preserve, and he knows or should know that it is relevant and material, he is not at liberty to destroy or dispose of it without notice to the defendant. If, for example, a car caught fire after having been struck from behind, and the remains of a gasoline can had been discovered in the trunk, we do not think a plaintiff could dispose of that evidence and only keep the remains of the car. The second tank was next to the grill when Allstate’s investigator saw it. As Thoreau once said, some circumstantial evidence is very strong, as when you find a trout in the milk. 5 Allstate would be quite surprised if it were told that it was required to act as if its insurers were always telling the truth, or that eyewitness testimony could not be disproved by circumstantial evidence. When Allstate’s investigator viewed the fire scene, he knew he was preserving evidence for a possible claim against Sunbeam. Since Sunbeam was not notified of the accident and given an opportunity to perform its own investigation, Allstate was required to preserve all reasonably relevant evidence, not just evidence that might support its claim. A reasonable investigator would have known that (1) a tank containing propane, if heated, will release highly flammable propane gas; (2) if the tank is ■ overfilled, heat from the ambient air and a nearby operating gas grill can be enough to cause it to do so; (3) the vented propane could have been drawn into the grill casting and ignited; and (4) soot and charring on the second tank indicate that it was touched by flames. Un *1279 der the circumstances, it was unreasonable of Allstate to have relied on statements by the insureds, (Janak Dep. at 36), that the tank was both empty and stored at a distance from the grill. We find that Sunbeam has been irremediably prejudiced. That prejudice cannot be overcome by a limiting instruction. If the jury were to be instructed that if it were to find that the second tank had been stored in close proximity to the grill, it must find that the second tank, not a defect in the grill, caused the fire, Sunbeam has nonetheless been deprived of what might have been convincing evidence that, in fact, the second tank was near the grill. The grill frame, described by the plaintiff, might have had marks indicating that flames came from somewhere other than the operating tank. The second tank could have had molten aluminum on top of it, establishing that it had been under the aluminum grill casting. Without these, which a reasonable investigator would have preserved, Sunbeam’s defense has been seriously and materially weakened. Accordingly, in the face of the destruction of this material evidence, we recommend that the complaint be dismissed. 6 By reason of plaintiffs act in destroying material evidence, we conclude that dismissal of the ease is an appropriate sanction deserving of Allstate’s cavalier attitude in this case. Knowing full well that subrogation efforts towards Sunbeam were a distinct possibility, Allstate nonetheless destroyed all evidence available for Sunbeam to formulate a cause and original analysis, and present that analysis to the fact finder. Sunbeam was inextricably harmed in its ability to defend the lawsuit. Accordingly, plaintiffs conduct requires, as an appropriate sanction, that the complaint be dismissed. The courts have favored this type of sanction under the circumstances of this case. Marrocco v. General Motors Corp., 966 F.2d 220 (7th Cir.1992). VIII. CONCLUSION For the above stated reasons, it is hereby recommended that Sunbeam’s motion for sanctions for evidence spoliation be granted, and that the complaint be dismissed with prejudice, and with all costs to defendant. Respectfully submitted, DATE: June 23, 1994. Footnotes [1] We assume Sunbeam refers to "Spoliation,” meaning the destruction of evidence so as to obstruct justice. See Black’s Law Dictionary, at 1257 (5th Ed.1979). [2] In a letter to the court dated November 19, 1993, Stewart D. Stoller, counsel for the Garcias, stated that the videotape was taken on July 4, 1990 at approximately 7:30 p.m. — the fire had started approximately 11:00 a.m. that morning. This letter was in response to the court’s directions for this information to be placed in the record. (Tr. 467). [3] Allstate objected to the admission of the photographs it had taken because they were not authenticated as reflecting the scene at the time of the fire. They were admitted, however, to show the scene as of the time they were taken, with the court to determine the weight they would be given as evidence of the appearance of the scene at the time of the fire. [4] Interestingly, Allstate chose not to produce an expert of its own who would offer expert opinions challenging the opinions of Sunbeam’s expert. [5] Thoreau, Henry David, Journal, November 11, 1850. [6] Because we find that the only appropriate sanction is the dismissal of the suit, this opinion is in the form of a Report and Recommendation pursuant to 28 U.S.C. § 636(b)(1)(B) rather than a ruling on sanctions pursuant to 28 U.S.C. § 636(b)(1)(A).