Full opinion text
OPINION HUDSON, Justice. This is an appeal from a suit arising out of an explosion at an underground salt-dome storage facility and pipeline pumping station. Twenty-two plaintiffs sued appellants, Seminole Pipeline Company, MAPCO, Incorporated, and Mid-America Pipeline, for damages resulting from the explosion. On appellants’ motion, the trial was bifurcated. Phase one of the trial resolved liability issues on actual and punitive damages, and determined the amount of the plaintiffs compensatory damages. The jury found each of the defendants (1) was negligent; (2) created a nuisance by their use of the property; (3) was grossly negligent; (4) acted with malice; (5) proximately caused the plaintiffs’ damages by their negligence; and (6) created a permanent nuisance. Also finding appellants jointly and severally liable for the total amount of plaintiffs’ damages, the jury awarded appellees approximately 5.4 million dollars in actual damages for property damage and mental anguish. In phase two, the jury assessed punitive damages and awarded appellees 46 million dollars in exemplary damages against each appellant. After several post-verdict hearings, the trial court reduced the exemplary damage award from 46 million dollars to four times the amount of actual damages, or approximately 21.6 million dollars. The trial court rendered judgment and the defendants appealed, raising forty-six points of error. A number of plaintiffs filed cross-appeals. On February 27, 1998, appellants and ten of the appellees, Pierre Cavaness, Carolyn Cavaness, Betty Crawford, Individually and as Trustee of the Betty A. Crawford Trust, Robert Hardie, James Jeffress, Sonya Jef-fress, Charles E. Nitsch, James A. Nitsch, Lois B. Nitsch, and David Schulz, filed a joint motion to dismiss all claims between them on appeal. We granted their motion and issued an opinion on March 26,1998. On March 27,1998, appellants and six appellees, Broad Leaf Partners, Ltd., Herbert Klatte, Esther Mae Klatte, Roy White, Dorothy White, and Vicki Schulz, also filed a joint motion to dismiss. On April 30, 1998, this court granted the motion and ordered all points of error asserted by appellants against the settling appellees and all cross-points of error asserted by the settling appellees dismissed. Twenty-two of appellants’ points of error remain for our consideration. Appellants contend (1) the evidence is legally and factually insufficient to support the jury’s finding of negligence and gross negligence, (2) the trial court misapplied the statutory cap on punitive damages and such damages are excessive, (3) the award for mental anguish is not supported by legally or factually sufficient evidence and is excessive, (4) the issue of mental anguish was improperly submitted to the jury, and (5) the trial court erred in calculating the amount of prejudgment interest. Appellees, W.W. O’Donnell, Regina O’Donnell, W.H. Tonn, III, and Jeanne Tonn, as Executrix of the Estate of W.H. Tonn, II, bring seven cross-points. We affirm the judgment of the trial court in part and also reverse and remand in part. I. CONSTRUCTION AND HISTORY OF THE BRENHAM CAVERN As common carrier pipeline companies, appellants transport ethane, propane, butane and other natural gas liquids (also known as NGL or LPG) through their pipeline networks. These products are actually gases, but they are most efficiently transported and stored while in a liquid state. To maintain these products in liquid form they must either be compressed or cooled. If they are not pressurized, these liquids will quickly revert back into a gaseous state at normal temperatures. To enhance their ability to deliver an uninterrupted supply of natural gas liquids, appellants have constructed a series of underground reservoirs into which they can divert and maintain these products under pressure. When appellants’ supply of natural gas liquids exceeds demand, the excess product is diverted into an underground cavern. Likewise, when demand exceeds the supply, appellants make up the difference by drawing from their reserves. Appellants began construction of the Bren-ham saltdome cavern storage facility in 1980, by a procedure known as “washing.” Fresh water was injected into an underground salt formation and the resulting brine was extracted. By this method, appellants “washed out” a huge subterranean cavern. In 1981, appellants obtained a permit from the Texas Railroad Commission to use the cavern as a storage facility for up to 150,000 barrels of natural gas liquids. To/From Brine Pond Typical Liquid Product Storage Normally, natural gas liquids stored in saltdome caverns “float” on an underground lake of heavy brine where they are safely contained under pressure. When additional natural gas liquids are injected into the cavern, brine is displaced and forced to the surface through a pipe known as the “brine string.” The excess brine is collected in a large surface pit. When natural gas liquids are extracted from the reservoir, brine from the surface pit refills the cavern and keeps the remaining product under pressure. Thus, while the product inventory may vary from day to day, there remains an unbroken “string of brine” linking the brine pit on the surface with the brine lake at the bottom of the subterranean cavern. The weight of the brine within the brine string maintains pressure on the cavern and prevents the stored natural gas liquids from vaporizing and escaping to the surface. If incoming product pushes the brine level below the bottom of the brine string, natural gas liquids will flow through the brine string to the surface. As the column of brine is displaced by product in the brine string, the pressure in the reservoir falls, natural gas liquids within the brine string and reservoir vaporize, and a “geyser” of escaping gas spews from the brine pit. As the pressure exerted on the product lessens, the flow of gas increases, and, if the path to the surface is not closed by mechanical means, all of the product inventory in the cavern will vaporize and escape into the atmosphere. When mixed with oxygen, NGL vapor is extremely flammable and capable of explosive ignition. Because of this potential hazard, the cavern operator must carefully monitor the facility’s inventory to insure that it never exceeds the capacity of the reservoir. To prevent product escaping in the event of an accidental overflow, appellants equipped the brine string at the Brenham cavern with a pressure sensing device called a Barksdale switch. Theoretically, if gas escaped through the brine string, the Barks-dale switch would detect the drop in pressure and activate a trip check system. Safety mechanisms would then automatically seal the cavern and prevent additional vapor from escaping into the air. The facility also had between 20 and 25 hazardous gas sensors around the facility to detect any escaping vapor. These sensors were connected to a trip check alarm inside the facility’s office. Appellants also installed monitors to detect changes in cavern pressure, and to indicate the volume of product entering or leaving the cavern. Moreover, the hazardous gas sensors were monitored 24 hours a day in Mid-America’s home office in Tulsa, Oklahoma by means of a computer network. From 1981 until the date of the explosion, Seminole periodically increased the facility’s storage capacity by “washing” the cavern with fresh water and further eroding the salt dome. Brine that could not be stored in surface holding pits was sold to outside vendors and transported off site. By November of 1991, Seminole had increased the Brenham cavern’s storage capacity from 150,000 to 336,600 barrels. Seminole informed the Texas Railroad Commission of its new storage capacity and set the working capacity of the cavern at 300,000 barrels, 90% of the actual calculated capacity. On March 31, 1992, Troy Boisen, a Mid-American employee living in Brenham, conducted a routine check of the trip check system, and found it to be operating normally- II. The Accident On April 7, 1992, at approximately 6:09 a.m., a hazardous gas alarm sounded at Mid-America’s office in Tulsa, Oklahoma. At 6:10 a.m., Dwight Aryain, the dispatcher monitoring the facility, called Boisen and informed him that a “hazardous gas fault” alarm had been activated at the cavern. Aryain asked Boisen to drive to the facility to determine the source of the signal. Boisen arose, took a shower, dressed, and stopped at a convenience store for a soft drink. He arrived at the cavern approximately 20 minutes after receiving the call. Boisen immediately observed that a large cloud of “fog” had engulfed the facility. When Boisen attempted to turn off the engine of his diesel pickup, the motor continued to run on the airborne gas vapor. Concluding there was probably not enough oxygen in the air to sustain him long enough to get to the facility’s manual shut-off valve, Boisen retreated to the house of some nearby residents. From there, Boisen telephoned the Tulsa dispatch office and advised it of the hazardous gas situation. He then called his supervisor, Ron Ritchey. Ritchey ordered Boisen to evacuate the area, retreat to a safe distance, and not to enter the vapor cloud. Boisen instructed the residents of the house to leave the area. Disregarding his personal safety and Ritchey’s explicit instructions, Boisen entered the vapor cloud and attempted to reach the main facility with the intention of shutting off the gas. Boisen encountered two other Seminole employees at the entrance of the facility, and the three men attempted to reach the main shut-off valve. The concentration of gas, however, was too great, and the men were forced to abandon the effort because the safety trailer which housed their emergency breathing apparatus and equipment necessary to perform the shut down was ninety miles away in Sugar Land. Shortly thereafter, Boisen noticed the headlights of a car disappear into the vapor cloud. Seconds later, the gas cloud exploded with the estimated force of a three-Mloton bomb. III. MAPCO’s Liability The jury found the explosion was proximately caused by the negligence of all three defendants. Mid-America and Seminole do not contest this finding. MAPCO, however, claims it was not negligent, and in its eleventh point of error, contends the evidence is legally and factually insufficient to support the jury’s finding in this regard. While Mid-America and Seminole concede they were responsible for the overfill, MAPCO claims it did not own or control the Brenham cavern, and it denies all liability for the incident. Appellees, on the other hand, argue that MAPCO possessed sufficient control over the construction and operation of the Brenham facility to permit the jury to infer its negligence under the doctrine of res ipsa loquitur. When both legal and factual sufficiency challenges are raised on appeal, the court must first examine the legal sufficiency of the evidence. See Glover v. Texas Gen. Indent. Co., 619 S.W.2d 400, 401 (Tex.1981) (per cu-riam). In reviewing a challenge to the legal sufficiency of the evidence, an appellate court considers only the evidence and inferences tending to support the trial court’s findings and disregards all evidence and inferences to the contrary. See Weirich v. Weirich, 833 S.W.2d 942, 945 (Tex.1992). Res ipsa loquitur is applicable only when two factors are present: (1) the character of the accident is such that it would not ordinarily occur in the absence of negligence; and (2) the instrumentality causing the injury is shown to have been under the management and control of the defendant. See Haddock v. Arnspiger, 793 S.W.2d 948, 950 (Tex.1990). Here, the record indicates MAP-CO’s direct involvement in the day-to-day operations of the cavern was extremely limited. Although MAPCO employees did perform occasional inspections of the facility inventory sheets, we do not find this fact alone to be the type of control necessary to impose liability under the doctrine of res ipsa loqui-tur. Appellees also attempt, however, to blend the doctrines of alter ego and res ipso loquitur and contend that MAPCO’s corporate structure and operations were so intertwined with those of Mid-America and Seminole that it effectively “controlled” cavern operations. For example, MAPCO is the sole shareholder of MAPCO Natural Gas Liquids Incorporated. MAPCO NGL, in turn, owns all the stock of Mid-America Pipeline Company. The senior vice-president of MAPCO, NGL, Robert Thomas Cronk, Jr., also serves as the senior vice-president of Seminole Pipeline Company. Mid-America supplies all of Seminole’s employees and owns 80% of its stock. Fred Isaacs, President of Seminole, is paid directly by MAPCO. Isaacs also admitted he was responsible for the construction of the Bren-ham cavern and its safety controls. In other words, Mid-America is the business entity that conducts MAPCO’s inter-state business and Seminole is the business entity which conducts MAPCO’s intra-state business in Texas. All three appellants are Oklahoma-based corporations located at the same address. As a general rule, a parent corporation is not liable for torts committed by its subsidiaries. See Lucas v. Texas Indus., Inc., 696 S.W.2d 372, 374 (Tex.1984). However, we will disregard the corporate fiction (1) when the fiction is used as a means of perpetrating fraud; (2) where a corporation is organized and operated as a mere tool or business conduit of another corporation; (3) where the corporate fiction is resorted to as a means of evading an existing legal obligation; (4) where the corporate fiction is employed to achieve or perpetrate a monopoly; (5) where the corporate fiction is used to circumvent a statute; or (6) where the corporate fiction is relied upon as a protection of crime or to justify wrong. See Castleberry v. Branscum, 721 S.W.2d 270, 272 (Tex.1986). Here, the record demonstrates the three defendant corporations were closely tied to one another, but a subsidiary corporation will not be regarded as the alter ego of its parent merely because of stock ownership, a duplication of some or all of the directors or officers, or an exercise of control that stock ownership gives to stockholders. See McFee v. Chevron Int’l Oil Co., Inc., 753 S.W.2d 469 (Tex.App.-Houston [1st Dist.] 1988, no writ). There must be something more than mere unity of financial interest, ownership and control before we will make the parent hable for the subsidiary’s tort. See Lucas, 696 S.W.2d at 374. In short, absent a showing of wrongdoing on the part of the parent corporation, Texas courts have refused to make the parent liable for its subsidiary’s torts. See Lubrizol v. Cardinal Constr., 868 F.2d 767, 771 (5th Cir.1989). Accordingly, we reject appellees contention that MAPCO’s corporate structure demonstrates it “controlled” cavern operations. Appellees contend, however, that appellants’ attorneys made certain admissions regarding the cause of the explosion. They claim almost no effort was made during trial to distinguish MAPCO from Mid-America or Seminole. In fact, all three defendants were represented by the same counsel and during voir dire, appellants’ lead trial counsel introduced everyone on the defense team and made the following comments during his examination of the venire: Immediately after this accident — this explosion — we really didn’t think that it was our responsibility. But as events unfolded and as the examination of what occurred did ensue, we have determined that it was caused by an overfill of the cavern. So whatever evidence is brought in here to determine that there was an overfill, it is really not addressing an issue that’s in contest here. We overfilled the cavern. Gas escaped from the cavern and it was ignited, and there was this terrible explosion. That explosion that we regret. We didn’t mean to do wrong. But unfortunately, we did. ... [T]he explosion came from our cavern. We have learned our lesson. We have indicated in this Court that we were the parties responsible for that overfill.... We have realized our mistakes. And in this case, we are ready to pay for the damages that we have caused. This theme reappeared during defense counsel’s closing statements to the jury when he said: When we were in court before on opening statement, I told you what our story— what our theme — was that we wanted to get over to you. It was simply this: That we didn’t mean to do wrong, but unfortunately we did. We learned our lesson, changed our ways, but weren’t given a chance to keep the cavern open. We accept our mistakes and we are ready to pay for damage we caused. But the plaintiffs want us to pay for things we didn’t do. And that’s why we think the plaintiffs’ requests for damages are unfair. MAPCO argues that counsel’s statements were “merely a recognition that the cavern was overfilled,” and not an admission of liability. Citing to Carroll Instrument Co. v. B.W.B. Controls, Inc., 677 S.W.2d 654, 659 (Tex.App.-Houston [1st Dist.] 1984, no writ), they also argue that MAPCO’s counsel did not make any binding judicial admissions and that if he did, Royce M. Parr, General Counsel and secretary for MAPCO NGL, unambiguously refuted those statements. Parr testified that he was present at trial, not as a company representative, but as an attorney on behalf for MAPCO Natural Gas Liquids, Inc., Mid-America Pipeline Company, and Seminole Pipeline Company. When asked whether the companies he was representing “admitted negligence for the damages sustained by [plaintiffs],” Parr responded “They have not.” The record also reveals that appellants’ counsel’s remarks during voir dire were prefaced with, “Let me explain Seminole’s position.” (Emphasis added). Likewise, counsel’s closing argument made reference to his opening statement where, again, his comments seem to have been directed toward Seminole. Before admitting “we” overfilled the cavern, counsel told the jury about the 36,000 barrel “safety factor” and said: Now that was the thinking of Seminole, that they were providing a cushion. That’s what was in their mind. That’s what they were conscious of. The truth is, the day before the explosion Seminole had no idea that they were that close or that they were anywhere near the safety — safe working capacity of the cavern. (Emphasis added). A judicial admission results when a party makes a statement of fact which conclusively disproves a right of recovery or defense he currently asserts. See H.E. Butt Grocery Co. v. Pais, 955 S.W.2d 384, 389 (Tex.App.-San Antonio 1997, no writ) (citing Gevinson v. Manhattan Constr. Co. of Okl., 449 S.W.2d 458, 466 (Tex.1969)). A judicial admission must be (1) made in the course of a judicial proceeding; (2) contrary to an essential fact for the party’s recovery [or defense]; (3) deliberate, clear and unequivocal; (4) related to a fact upon which judgment for the opposing party could be based; and (5) enforcing the admission would be consistent with public policy. See Sepulveda v. Krishnan, 839 S.W.2d 132, 135 (Tex.App.-Corpus Christi 1992), aff'd, 916 S.W.2d 478 (Tex.1995). Counsel’s statements on behalf of his client will be deemed judicial admissions only if they satisfy the five criteria listed above. See Sepulveda, 839 S.W.2d at 135; see also Shafer v. Bedard, 761 S.W.2d 126, 129-30 (Tex.App.-Dallas 1988, no writ); Hochmetal Africa (PTY), Ltd. v. Metals, Inc., 566 S.W.2d 715, 718 (Tex.Civ.App.-Corpus Christi 1978, no writ); Rosse v. Northern Pump Co., 353 S.W.2d 287, 292 (Tex.Civ.App.-Austin 1962, writ ref d n.r.e.). The principle of binding a party by his judicial admission should be applied with caution because the effect is that the declarant swears himself out of court. See Esteve Cotton Co. v. Hancock, 539 S.W.2d 145, 157 (Tex.Civ.App.-Amarillo 1976, writ ref'd n.r.e.). The statements made here were, at best, equivocal. We find the character of the statements do not rise to the level of “judicial admissions.” In the event that MAPCO’s liability cannot be based upon judicial admissions, appellees contend it can be supported by the jury’s finding of a permanent nuisance because MAPCO does not challenge the finding. The jury, however, only assessed damages stemming from the “occurrence in question.” We find the only reasonable interpretation of this phrase has reference to the explosion. While a permanent nuisance might fairly be described as a “condition,” it is rarely considered an “occurrence.” Accordingly, there is no evidence to support the conclusion that MAPCO was negligent or that it possessed such control over the operation of the Brenham facility to render it liable for damages under the res ipsa loquitur doctrine. Appellants’ eleventh point of error is sustained. IV. Gross negligence and malice The jury also found all three appellants acted with gross negligence and malice. These findings provide the basis for the award of exemplary damages. In their first six points of error, appellants attack these findings, and hence the exemplary damages, by arguing the trial court erred in overruling their motions for (1) directed verdict, (2) judgment n.o.v., and (3) new trial because the evidence is legally and factually insufficient to support any finding of gross negligence or malice as to each of the appellants. Claiming appellees’ own experts admit appellants had no subjective awareness of the risk of explosion, appellants argue appellees’ “theory of gross negligence seems to be that an abundance of evidence of simple negligence combined with a tragic result meta-morphosizes the appellant’s conduct into gross negligence.” Relying upon Transportation Ins. Co. v. Moriel, appellants contend that evidence of ordinary negligence, though it may lead to massive damage, does not constitute proof of gross negligence. See 879 S.W.2d 10, 21-22 (Tex.1994). Appellants also argue that even if some evidence shows they could have taken additional safety precautions or used safer equipment to lessen the risk of injury, such evidence does not necessarily establish gross negligence. See Sheffield Div., Armco Steel Corp. v. Jones, 376 S.W.2d 825 (Tex.1964); Universal Servs. Co., Inc. v. Ung, 904 S.W.2d 638, 641 (Tex.1995). Where the cause of action accrues, as it did here, before September 1, 1995, a punitive damage award requires proof that the damage or harm resulted from fraud, malice, or gross negligence. In such cases, the Civil Practice and Remedies Code defines gross negligence as: [M]ore than momentary thoughtlessness, inadvertence, or error of judgment. It means such an entire want of care as to establish that the act or omission was the result of actual conscious indifference to the rights, safety, or welfare of the persons affected. See Act of June 3, 1987, 70th Leg., 1st C.S., ch. 2, § 2.12, 1987 Tex. Gen. Laws 44 (formerly codifed at Tex. Civ. PRAC. & Rem.Code § 41.001(5)); see also Burk Royalty Co. v. Walls, 616 S.W.2d 911, 920 (Tex.1981). In Moriel, the Supreme Court held that gross negligence involves two components: (1) the defendant’s act or omission and (2) the defendant’s mental state. See 879 S.W.2d at 21. Gross negligence differs form ordinary negligence in two respects: (1) the defendant must be “consciously indifferent,” and (2) his conduct must “create an extreme degree of risk.” Id. The test for gross negligence contains both an objective and subjective component. Id. at 21-22. Subjectively, the defendant must have actual awareness of the extreme risk created by his conduct. Id. at 22. Objectively, the defendant’s conduct must involve an “extreme degree of risk,” a threshold significantly higher than the objective “reasonable person” test for negligence. Id. Extreme risk is a function of both the magnitude and the probability of the anticipated injury to the plaintiff. Id. An act or omission that is merely thoughtless, careless, or not inordinately risky cannot be grossly negligent. Id. For actions accruing prior to September 1, 1995, the Civil Practice and Remedies Code also defines malice as “an act that is carried out by the defendant with a flagrant disregard for the rights of others and with actual awareness on the part of the defendant that the act will, in reasonable probability, result in human death, great bodily harm, or property damage.” A. Standard of Review In reviewing the evidence under a no evidence point, we consider all the evidence in the light most favorable to the prevailing party, indulging every reasonable inference in that party’s favor. See Associated Indem. Corp. v. CAT Contracting, Inc., 964 S.W.2d 276, 285-86 (Tex.1998); see also Moriel, 879 S.W.2d at 24. In evaluating legal sufficiency, we are required to determine whether the proffered evidence as a whole rises to a level that would enable reasonable and fair-minded people to differ in their conclusions. Moriel, 879 S.W.2d at 25. The Supreme Court has held that when conducting a factual sufficiency review of a punitive damages award, we must detail the relevant evidence, explaining why that evidence either supports or does not support the punitive damages award in light of the factors set out in Alamo Nat’l Bank v. Kraus, 616 S.W.2d 908, 910 (Tex.1981). See Moriel, 879 S.W.2d at 31. Therefore, in reviewing the appropriateness of the jury’s award we will consider (1) the nature of the wrong; (2) the character of the conduct involved; (3) the degree of culpability of the wrongdoer; (4) the situation and sensibilities of the parties affected by the wrong; and (5) the extent to which such conduct offends a public sense of justice and propriety. See Kraus, 616 S.W.2d at 910. B. Subjective awareness and conscious indifference While appellants freely admit that neighboring property owners were damaged when the cavern facility exploded, they strenuously contend there was little, if any, evidence to support a finding of gross negligence or malice. Because they had numerous sáfety mechanisms and procedures in place at the time of the explosion, appellants claim the jury had no basis for finding they were subjectively aware of the risk, yet consciously indifferent to the rights, safety, or welfare of others. Appellants point to evidence that the cavern had a capacity of 336,600 barrels and the anticipated inventory on the day of the accident was only 288,305 barrels, or 48,000 barrels less than cavern capacity. Moreover, the cavern had been filled with 297,649 barrels of natural gas liquids without incident on March 11, 1992. Accordingly, on April 7, 1992, appellants believed the cavern contained 9,344 fewer barrels than it did on March 11, 1992. Further, they tested the Barksdale switch a week before the explosion and found it to be fully operational. The technician who ran the test found the trip cheek system,’ solenoid, tubing, plunger, valve, and valve handle all functioning normally. He also recalibrated the Barksdale switch at that time and set it to activate if the pressure in the brine string fell below 100 psi. Finally, appellants contend their confidence in the safety system was reasonable because the same trip check system had worked in 1988 and had closed in the cavern when natural gas liquids began escaping through the brine string. Appellants argue that this evidence shows them to be subjectively justified in believing their safety mechanisms and procedures would shut down the cavern in the event of an overflow. 1. Awareness of the risk We begin our analysis by noting that the transportation and storage of natural gas is an inherently hazardous venture because the potential for disaster is enormous. The Supreme Court has observed: In the operation of its business a. gas company is bound to exercise such care and diligence as to avoid injury to the ... property of others by the escape of gas.... The degree of care required to prevent injury has been variously stated to be ordinary care, due care, due and reasonable care, great care, a high degree of care, and the highest degree of care. This variety of expression means no more, however, than that care and diligence should always vary according to the exigencies which require vigilance and attention, conforming in amount and degree to the particular circumstances under which they are to be exerted.... Therefore, in view of the highly dangerous character of gas and its tendency to escape, a gas company must use a degree of care to prevent damage commensurate to the danger which it is its duty to avoid. Prudential Fire Ins. Co. v. United Gas Corp., 145 Tex. 257, 199 S.W.2d 767, 772-3 (1946). Appellants were well aware of the potential danger for property damage, personal injury, and death that could result if the cavern was overfilled. Wes Baker, operations supervisor for Seminole Pipeline from 1981 to 1985, testified that Seminole’s instructor’s manual on meter tickets and daily operating reports for Mid-America contained the following admonition: Accurate product inventory records may help prevent accidents resulting in overfilling a cavern, tank or section of pipeline. Overfilling and/or overpressure of a vessel can result in personal injury, property damage or worse. (Emphasis added). Charles Vest, an employee of Mid-America, testified that the Y-grade product which was stored in the cavern at the time of the explosion had a typical vapor pressure of 250 pounds. A Mid-America publication states: If the pressure on Y-grade is reduced to or below it’s [sic] vapor pressure, the product will suddenly and very rapidly vaporize. This is called flashing. Flashing product can be very dangerous. Appellants were also aware that natural gas liquids will enter the brine string if the cavern is overfilled, and if natural gas liquids entered the brine string, cavern pressure would be reduced and flashing would occur. Thus, the first line of defense against flashing is to make certain the cavern’s inventory of natural gas liquids does not exceed its capacity. a. Inaccuracy of inventory estimates While appellants used meters to track the number of barrels of product moving in and out of the cavern, Wes Baker, an operations supervisor, testified that variables such as temperature and pressure would induce errors in any meter. Because meters cannot precisely measure the amount of product passing in or out of the cavern, it was impossible for appellants to know the exact number of barrels of natural gas liquids stored in the cavern. Although errors induced by inaccurate meters may be relatively small, this inaccuracy is compounded each time product is added to or subtracted from inventory. Only when all natural gas liquids are removed and the cavern is filled with brine alone, can the product inventory be precisely known, i.e., zero barrels. The record shows that appellants were subjectively and objectively aware that as they added and removed product without emptying the cavern, their inventory estimates became less reliable. The longer appellants waited between emptying the cavern, the greater the chance of a major error in inventory calculations. During normal operations, gas was continuously diverted in and out of a series of caverns along appellants’ pipeline network. In the normal course of business, appellants emptied their caverns approximately once a year to insure the accuracy of their inventory estimates. However, because the Brenham facility sat atop a large cavern, Wes Baker testified that appellants attempted to empty it more frequently than some of the other reservoirs. In fact, appellants’ records indicate they emptied the Brenham facility and “re-zeroed” the inventory base line on an average of 16 times per year from 1982 through 1991. Curiously, appellants abandoned this practice in 1991. During the first seven months of 1991, the cavern was emptied eight times. However, between July 12, 1991, and April 7, 1992, the date of the explosion, the cavern was never emptied. In fact, this nine-month hiatus appears to be the longest such interval in the history of the facility. Because serious inventory discrepancies were often discovered each time the cavern was emptied, prudence required this procedure to be repeated frequently. For example, when appellants did not empty the cavern for four months between February 23, 1987, and June 27, 1987, they discovered an inventory discrepancy of 28,710 barrels. When appellants did not empty the cavern during a five-month period in 1989, they discovered they had underestimated cavern inventory by 24,347 barrels. In fact, significant errors occurred even during brief intervals. Records show the cavern was emptied on September 18, 1987; the inventory was reset to zero, and the cavern was refilled. Three days later, on September 21, 1987, the cavern was emptied again and found to contain 15,129 more barrels of natural gas liquids than anticipated. The cavern was refilled and emptied the following day. After only one day’s use, appellants discovered they had underestimated inventory by an additional 6,818 ban-els. The inventory base line was reset to zero, and the cavern was put back in use. When it was emptied two days later on September 24, 1987, appellants found an unexpected 9,953 barrels in the cavern. In other words, if the cavern had not been emptied three times during the six-day period between September 18, 1987, and September 24, 1987, the cavern would have contained 31,-900 more barrels of natural gas liquids than anticipated. Six months before the accident, in October of 1991, a data clerk in Tulsa discovered a mistake of 28,406 barrels had been made in Brenham. When Charles Vest discovered he had made a significant miscalculation regarding the amount of product transferred from the Brenham cavern into a customer’s line, he wrote a memorandum to his supervisor in which he admitted: I misunderstood the accounting procedure to be used in accounting for barrels delivered to UTTCO from ten days earlier this month. Due to this we subtracted barrels delivered to UTTCO from cavern inventory in error. The net result is that we need to add 28,406 barrels to the cavern inventory at Brenham. On March 11, 1992, less than a month before the accident, inventory estimates indicated the volume of the cavern had reached 297,649 barrels, an all-time high. Because the cavern had not been emptied in 8 months, this estimate was highly suspect and an employee was assigned to “sit” the cavern, i.e., remain at the site all night to watch for any signs of an overfill. The cavern did not overflow that evening. The cavern’s inventory was subsequently reduced and appellants deemed it unnecessary to continue to “sit” the cavern during the evening hours. At the end of March, eight days before the explosion, Seminole held a meeting to discuss cavern inventory. Because there had been a series of serious inventory miscalculations, Charles Vest suggested the cavern be emptied and the inventory “reset to. zero.” Vest’s suggestion, however, was disregarded, the cavern was not emptied, the inventory calibrations were not reset, and more natural gas liquids were pumped beneath the ground. David Allen Haney, scheduler for MAP-CO/Mid-Ameriea, also discussed the possibility of zeroing-out the cavern. He testified he told John Mobley, Vest’s field supervisor, that schedulers are totally dependent on the data provided by the field staff, and that just two days before the accident, he and Mobley “discussed that [the cavern] was up there at the operating limit.... We definitely talked about lowering the inventory.” Despite his concerns, the cavern was not emptied and Haney personally ordered approximately 36,-000 additional barrels placed in the cavern. After the accident, Wes Baker headed up a team to reconstruct the cavern’s inventory. The starting point for the reconstruction was the last time the inventory had been reset to zero, nine months earlier. The team estimated that, at the time of the accident, there were approximately 50,000 more barrels of product in the cavern than shown on appellant’s records. One thousand barrels of the overage was attributable to 85 different arithmetic errors made by the cavern’s operators. Almost 700 other errors resulted from operators using incorrect temperature and pressure correction factors in their calculations. Baker’s team finally determined that the size of the overage was an aggregate of mismeasurements made by the cavern operators over a 270-day time period. b. Inadequacy of the safety trip check system The trip check system was appellants’ second line of defense against a catastrophic failure. It was designed to shut in the cavern if gas began flowing up the brine string, and it did so successfully on one occasion several years prior to the explosion. This system, however, had a mixed history of success. On March 18, 1982, Charles Vest was in the station control building at the Brenham facility when a 20 to 30 foot plume of water and gas began spewing out of brine pit number one. Because the Barksdale switch did not shut in the cavern, Vest had to manually close the trip check valve. Vest reported the system failure to the operations supervisor, Wes Baker. Likewise, in 1988, Vest observed gas escaping to the surface of brine pit number one when the Barksdale switch failed to shut in the cavern during an inadvertent overfill. Vest was again forced to manually shut down the cavern. Fortunately, these incidents occurred during normal work hours; otherwise, the leaks would have continued unabated. Troy Boisen was primarily in charge of testing and calibrating the Barksdale switch every six months. If the switch did not operate properly three times in succession, he replaced it. Boisen testified that appellants had no written company procedures concerning the periodic Department of Transportation (DOT) inspections of the Barksdale switch. When performing inspections and maintenance on the switch, Boisen merely set the switch to the pressure range marked on the DOT report, 100 p.s.i. Admitting that he did not know the switch’s recommended activation range, Boisen was not aware the manufacturer’s specifications cautioned that if the switch’s settings were set beneath the recommended range, the switch would not operate reliably. Rudolf Wolf, president and general manager of Barksdale, Incorporated, the manufacturer of the Barksdale switch, testified that his company conducted a post-accident examination of the switch used at the Brenham saltdome facility. Barksdale’s technicians discovered the switch used by appellants was an odd assemblage of “cannibalized” parts taken from other Barksdale switches. The switch housing had been manufactured prior to November, 1984. The bourdon tube had been put in stock in January, 1989. The microswitch was manufactured in 1980, and the block bracket was made iij 1985. Sometime after the switch was shipped from the company, its two set screws had been either removed or loosened and retightened. One screw was set at 90 pounds and the other at 100 pounds. An adjustable bourdon tube’s switches typically range from 150 to 200 psi up to 18,000 psi. The adjustable range for the switch involved in this accident was designed to operate between 160 to 2,000 psi. When set at 100 psi, as it was here, Barks-dale’s literature indicates the switch would yield poor accuracy. Barksdale technicians attempted to “trip” the switch by applying pressure ranging from 10 to 200 psi. The switch did not activate. Further examination of the switch revealed corrosion inside the pressure connection, switch housing, and the threads of the switch housing. The surge dampener was totally obstructed with particles of brown material that could have been a combination of sand, rust, pipe scale, - and salt crystals. When technicians removed the surge dampener, the switch worked properly. Appellees’ expert witness testified that he had examined the switch and concluded that it had not been properly maintained. He based his conclusions on (1) corrosion inside the switch’s pressure connection; (2) corrosion in the switch housing; (3) corrosion in the threads of the switch housing; (4) dirt and debris clogging the sensing line; and (5) brown material totally blocking the surge dampener. 2. Conscious disregard of the danger Despite appellants’ knowledge that the Barksdale switch was the only automated, non-manual, safety device on the brine string capable of shutting in the cavern, appellants only tested its reliability semi-annually. Appellants also knew the switch had been rebuilt from used parts and had made the conscious decision to set it below its recommended operating range. The employee in charge of maintaining, testing, and calibrating the switch was, by his own admission, ignorant of the manufacturer’s recommended pressure range for this switch. Despite the sporadic reliability of them automated safety mechanisms, appellants left the Brenham cavern unattended 16 hours of each day. During the evening hours, the last line of defense was a monitoring station in Tulsa, Oklahoma, where four dispatchers watch over Mid-America’s approximately 10,000 miles of pipeline twenty-four hours a day. On the day of the accident, three dispatchers were monitoring the pipeline. That morning, Gary Dwayne Nabors, dispatcher for Mid-America, was supervising Dwight A. Aryain, a dispatcher trainee. Nabors, had undergone four months of training before he assumed the position of dispatcher. At no time during his training had he ever been placed in a mock situation involving a cavern overfill, nor did appellants possess any equipment to simulate an emergency scenario. When the Tulsa monitoring center received the hazardous gas alarm, Nabors and Aryain followed the company’s standard operating procedure. Aryain telephoned Troy Boisen and asked him to drive out to the Brenham cavern and “check it out.” Because only one sensor detected gas and the two men did not see anything on their system console that would indicate gas had entered the brine string, they chose not to notify emergency personnel. Aryain assumed that if the cavern had overfilled, the safety check system would shut down the cavern automatically. Although the hazardous gas alarm sounded 27 times, Aryain did not wait for Boisen to call back, but left the building when his shift ended. Aryain was relieved by Roger Collins, a dispatcher with fifteen years experience. Aryain and Norris briefed Collins on the gas alarm and told him that they had called Boisen, but did not inform him as to the time of the call. Collins confirmed that it was company policy not to treat one hazard sensor as an emergency situation. He had confidence in the minicomputer system, and testified that “[i]f everything worked like it should have there would have been enough information” to realize there was a dangerous situation in Brenham. Unfortunately, Collins’ monitors never registered a trip check, fire, rate of flow change, or any other faults. As Collins later related, “[there] simply wasn’t enough information there to tell me that we had a problem.” If the system had reported more faults, Collins said he would have alerted emergency personnel to the danger. However, even if Tulsa had realized there was an emergency situation in Brenham, the monitoring facility had no means by which it could remotely activate the trip check system or close in the cavern. The only way to stop product from flowing out of the cavern was for someone to physically enter the Brenham facility and manually close the valve. Alan Wolf, an employee of Mid-America, was, at the time of the accident, a computer technician maintaining the minicomputer system. He testified that the software for the system was developed in 1974 by TRW, and was installed in 1982 on two mirrored minicomputers. While this was considered “state of the art” in 1982, Wolf acknowledged that by 1990, it was obvious “we were pushing the limits of the system, as far as how many points it could handle, how fast it could process data.” The system had several other weaknesses, including a lack of software packages to detect leaks, and an unfortunate tendency to crash frequently. Wolf testified the system “was being rebooted every three to five days.” During a reboot, dispatchers did not have control of the system. Data from the minicomputer system was displayed in a tabular format instead of real-time graphic readouts, and the system did not have the ability to display real-time graphs of pressures, discharge pressures, and cavern pressure. Wolf also conceded the computer system in place at the time of the explosion did not have the capability to discern whether the cavern was being overfilled. Wolf admitted the system was limited in the length of the alarm message it could display, and at the time of the explosion, it was not possible to record data as it came in and replay past events. In February, 1992, less than a month before the accident, Wolf wrote a memo outlining new features he would like to see on the new system, including a simulator training station. Timothy Richard Moran, an employee of PB-KBB, Inc., an engineering construction company, was hired by MidAmerica, Seminole, and MAPCO, to investigate the accident. Moran concluded that having only one safety device on the brine string was not “a reasonable design” and further determined that having only one device on the brine string would not be what he considered “an adequate design.” He also stated that the explosion would not have occurred if the Brenham facility had installed and maintained a flare. Such a flare would have ignited the escaping gas long before it could accumulate into an explosive cloud. In a safety publication authored by Moran, he recommended (1) an excess flow detector on the product line; (2) redundant safety systems employing two devices (ESD valves) on the brine line; and (3) two devices on the product line set to detect different parameters, including pressure flow and/or conductance and pressure. When an abnormal situation occurred, one of the detectors would detect the anomaly and shut down the system. PB-KBB also recommended a separator/flare system. C. Gross negligence The testimony of appellants’ own employees establishes their subjective and objective awareness that their meters were imprecise and that their inventory estimates could differ from actual inventory by thousands or tens of thousands of barrels. Evidence established that appellants were also aware that (1) their inventory estimates were often in error; (2) they had grossly underestimated the cavern’s inventory in 1982; (3) a major calculation error had taken place six months earlier; (4) they had a “cloud” on the inventory; (5) the cavern’s inventory was reaching capacity; (6) the only way to get an accurate measurement of the cavern’s inventory was to zero out the cavern; and (7) it had been nine months since the cavern had been emptied. As early as 1990, appellants knew they were pushing the limits of their old computer monitoring system. They were aware (1) it had no packages in place that would detect leaks; (2) the system went down every three to five days; (3) when the system was rebooted dispatchers did not have control of the system; (4) data was displayed in a tabular format instead of real-time graphic readouts of pressures, discharge pressures, and cavern pressure; and (5) even if an emergency situation in Brenham were discovered, the dispatch facility had no capability to close the “trip check” remotely. We find the evidence is legally sufficient and conclude the jury was justified in finding Seminole and Mid-America had an actual subjective awareness of the risk of an overfill. They were also aware that an overfill would present a danger of a massive explosion. Nevertheless, they consciously proceeded to add natural gas liquids to the cavern without manning the facility, installing a flare system, setting up duplicate or redundant trip cheek safety mechanisms, or otherwise attempting to lessen the danger. In short, Seminole and Mid-America proceeded with conscious indifference to the rights, safety, and welfare of others. We also reject appellants’ contention that the evidence was factually insufficient to support the jury’s finding of gross negligence. We are aware that appellants did not deliberately overfill the cavern and once the leak was discovered, their employees risked their lives and suffered personal injury attempting to stem the flow of gas. Further, appellants attempted to maintain a safety margin of 36,000 barrels by never filling the cavern with more than 300,000 barrels of natural gas liquids. Finally, appellants were aware the Barksdale switch had successfully shut in the cavern on one occasion when it was inadvertently overfilled. However, Seminole and Mid-America were also aware the Barksdale switch had failed on at least two occasions. The Brenham facility was designed by the president of Seminole and appellants knew (1) the Barks-dale switch was the only safety device at the facility that had the capability of shutting in the cavern in the event of an overfill; (2) they had only the vaguest notion of the size of their product inventory because nine months had elapsed since the cavern had been emptied; (3) if the cavern overfilled during the night, no employees would be at the facility to visibly observe the escaping gas; (4) the computer system designed to alert company dispatchers in Tulsa was approaching obsolescence and was incapable of providing much of the information needed to discern whether the cavern had overfilled; (5) even if alerted to the danger, Mid-American dispatchers in Tulsa had no means of remotely shutting in the cavern; (6) by the time a local employee could be alerted and directed to manually shut in the cavern, he would likely need emergency breathing apparatus stored 90 miles from the facility. The jury’s finding of gross negligence is amply supported by the evidence, and we cannot say its finding is so contrary to the overwhelming weight of the evidence as to be clearly wrong and unjust. D. Malice “Gross negligence” and “malice” are not synonymous. Historically, “gross negligence” has always been perceived as a lower and less deliberate level of culpability than “malice,” and facts establishing the former will not necessarily support the later. The concepts of gross negligence and malice first appear in criminal law sometime after the thirteenth century as a means of distinguishing levels of homicide based upon a defendant’s mens rea. See 2 Sir FREDERICK Pollock and Frederic William Mait-land, The History of English Law 464-86 (2d ed. Cambridge 1899). The distinction between a willful murder and a homicide by misadventure was expressed in terms of ma-litia praemeditata and malitia excogitata. An intentional homicide, ie., one committed with “malice aforethought” was punishable by death, while an accidental homicide due to negligence carried no criminal liability. Later, an intermediate level of culpability, ie., “gross” or “criminal” negligence, was recognized in cases of involuntary manslaughter. See United States v. Browner, 889 F.2d 549, 553 (5th Cir.1989) (stating that under common law the “requisite mental state [for manslaughter] is reduced to ‘gross’ or ‘criminal’ negligence, a culpability that is far more serious than ordinary tort negligence but still falls short of that most extreme recklessness and wantonness required for ‘depraved heart’ malice”). “Malice” constituted the highest level of criminal culpability and was defined as the “willful and intentional doing of some wrongful act.” See Crutchfield v. State, 110 Tex. Crim. 420, 10 S.W.2d 119, 121 (1928). This is not to say, however, that the defendant had to intend a specific result of his actions or even hate his victim. In Banks v. State, where the defendant fired indiscriminately at a moving train and killed a man he did not know, the Court of Criminal Appeals wrote: Malice may be toward a group of persons as well as toward an individual. It may exist without former grudges or antecedent menaces. The intentional doing of any wrongful act in such manner and under such circumstances as that the death of a human being may result therefrom is malice. 85 Tex.Crim. 165, 211 S.W. 217 (1919). “Gross negligence,” while less culpable than “malice,” was also based upon conscious decisions and deliberate conduct. “The difference between malice, which will support conviction for murder, and gross negligence, which will permit of conviction only for manslaughter, is one of degree rather than kind.” United States v. Fleming, 739 F.2d 945, 948 (4th Cir.1984). Malice “contemplates a subjective awareness of a higher degree of risk than does gross negligence, and involves an element of wantonness which is absent in gross negligence.” People v. Watson, 30 Cal.3d 290, 179 Cal.Rptr. 43, 637 P.2d 279 (Cal.1981). The concepts of “gross negligence” and “malice” migrated into civil law as civil liability began to attach to criminal homicides. While early Teutonic law supplied a measure of compensation to the families of murder victims, the common law provided no recompense. Under the common law, a cause of action for personal injuries resulting in death terminated upon either the death of the victim, or the wrongdoer. See Burk Royalty, 616 S.W.2d at 916. This rule came under attack around the middle of the nineteenth century, and in 1869 the Texas Constitution added a provision for the recovery of exemplary damages in eases of homicide resulting from a “willful act or omission.” Id. In 1876, the Constitution expanded this right to include homicides due to gross negligence. In 1974, “malice” disappeared from Texas criminal law with the adoption of a new penal code. It survives, however, in civil law where it remains distinctly different from gross negligence. See Randolph v. Edmonds, 185 Tenn. 37, 43-45, 202 S.W.2d 664, 667 (1947). The primary distinction between “malice” and “gross negligence” is, in both civil and criminal law, the certainty of causing harm. When a defendant acts with gross negligence, his conscious indifference to the risk subjects the plaintiff to a “likelihood” of serious injury. See Mobil Oil Corp. v. Ellender, 968 S.W.2d 917, 921 (Tex.1998). However, when a defendant acts with malice, he is aware that his flagrant disregard of the plaintiffs rights will, “in reasonable probability,” result in death, great bodily harm, or property damage. See Act of June 3, 1987, 70th Leg., 1st C.S., ch. 2, § 2.12, 1987 Tex. Gen. Laws 1, 44 (formerly codified at Tex. Civ. PRAC. & Rem.Code § 41.001). Here, appellees presented considerable evidence that appellants were cognizant of the hazardous danger which would occur if the cavern was overfilled. As we have previously stated, the evidence further shows appellants knew their inventory estimates were suspect. Appellants, nevertheless, consciously disregarded the risk, thereby creating the “likelihood” of a catastrophic explosion and serious injury. However, we find no evidence in the record before us that appellants believed the cavern would probably overflow. In fact, the undisputed evidence shows appellants believed the cavern inventory was at least 9,000 barrels below what it had been less than a month before the accident. While inventory inaccuracies certainly raised a likelihood the cavern might overflow, there is no evidence to show appellants believed it would probably overflow. Accordingly, we hold the evidence is legally and factually sufficient to support the jury’s findings of gross negligence, but legally and factually insufficient to sustain the finding of malice. We therefore sustain appellants’ fourth point of error and overrule their first, second, third, fifth, and sixth points of error. V. Punitive Damages Cap In points of error seven through ten, appellants contend the trial judge misinterpreted the punitive damages cap found in Section 41.007 of the Civil Practice and Remedies Code. The aggregate compensatory damages found by the jury for all of the plaintiffs, including past and future mental anguish, was $5,395,400. The jury awarded punitive damages in the amount of $46,000,000 against each defendant. The jury then apportioned the punitive damages award among the various plaintiffs by assigning each a specific percentage of the award. The portion assigned to each of the O’Donnells, for example, was 14.99%. The portion assigned to W.H. Tonn and Jeanne Tonn was 5.79% and 4.74% respectively. Thus, the jury’s verdict for the appellees was: Jury’s verdict Plaintiff Actual 1 Damages Punitive Damages / each defendant W. W. O’Donnell $842,000 $6,895,400 Regina O’Donnell $842,000 $6,895,400 W. H. Tonn, III $353,250 $2,663,400 Jeanne Tonn $294,250 $2,180,400 The statutory cap on punitive damages in effect at the time this cause of action accrued provides that “exemplary damages awarded against a defendant may not exceed four times the amount of actual damages or $200,-000, whichever is greater.” Attempting to implement the cap, the trial judge reduced the punitive damages to $21,581,600 per defendant (or four times the total actual damages of $5,395,400). The court then apportioned this sum according to the percentages devised by the jury as follows: Court’s judgment Plaintiff Actual Punitive Damages / Damages each defendant W. W. O’Donnell $842,000 $3,235,081.84 Regina O’Donnell $842,000 $3,235,081.84 W. H. Tonn, III $353,250 $1,249,574.64 Jeanne Tonn $294,250 $1,022,967.84 Appellants complain that when the trial court used the total actual damages figure assessed against all of the defendants to calculate the amount of punitive damages owed by each defendant, it effectively tripled the capped award. Instead, appellants contend the trial court should have used the following equation to calculate the statutory cap: (actual damages) x 4 = total punitive damages allowed However, the trial court relied upon the following formula: (actual damages) x 4 x (number of defendants) = total punitive damages allowed Appellants claim the trial court has improperly capped the punitive damages at twelve times actual damages instead of four times actual damages. A. Application of the cap We must decide whether the damage cap applies to the plaintiffs’ award or the defendants’ liability. If we find the cap should be applied so as to limit each plaintiff’s award of punitive damages to four times his actual damages, the trial court’s judgment violates the statute. However, if we find the cap is applied to limit each defendant’s liability to four times the plaintiffs actual damages, then we must affirm trial court’s judgment as proper. When construing a statute, we are to consider the “object sought to be attained.” See Tex. Gov’t Code Ann. § 311.023 (Vernon 1988). It seems apparent that one of the purposes of the 1987 tort reform legislation was to make the Texas civil justice system more predictable. See John T. Montford and Will G. Barber, 1987 Texas Tort Reform: The Quest for a Fairer and More Predictable Texas Civil Justice System, 25 Hous. L.Rev. 59, 69 (1988). Presumably, the cap found in Section 41.007 was to provide a measure of predictability by narrowing the range of punitive damages that might be awarded to a plaintiff. However, Senator John T. Mont-ford, the principal author of the statute, has said Section 41.007 was intended to apply on a “per defendant” basis: Under section 41.007, a claimant cannot recover exemplary damages from any one defendant in excess of four times the amount of the claimant’s actual damages, but can recover up to that amount from each defendant against whom the trier of fact awards exemplary damages. Montford, 25 Hous. L.Rev. at 335 (emphasis added). Ironically, if we find the statute applies on a “per defendant” basis, it will have little value as a “cap.” In eases involving multiple tort-feasors, it is not difficult to imagine hypothetical scenarios where the cap would be completely