Kalamazoo River Study Group, Plaintiff-appellant, v. Rockwell International Corporation and Eaton Corporation, Defendants-appellees, 355 F.3d 574 (6th Cir. 2004)

Annotate this Case
US Court of Appeals for the Sixth Circuit - 355 F.3d 574 (6th Cir. 2004) Argued September 9, 2003
Decided and Filed January 14, 2004

Appeal from the United States District Court for the Western District of Michigan, Robert Holmes Bell, Chief Judge. COPYRIGHT MATERIAL OMITTED COPYRIGHT MATERIAL OMITTED Alan C. Bennett (briefed), Law, Weathers & Richardson, Grand Rapids, MI, Jerome T. Wolf (argued and briefed), James L. Moeller (briefed), Amy E. Bauman, David S. Ladwig, Katherine I. Funk, Sonnenschein, Nath & Rosenthal, Kansas City, MO, for Appellant.

Kathryn J. Humphrey (briefed), Dykema Gossett, Detroit, MI, Joseph C. Basta (argued and briefed), Dykema Gossett, Ann Arbor, MI, for Appellees.

Before MOORE and GILMAN, Circuit Judges; MILLS, District Judge.* 

OPINION

MOORE, Circuit Judge.


In 1990, federal and state environmental authorities officially recognized the massive polychlorinated biphenyls ("PCB") contamination of the Kalamazoo River in Michigan by placing a portion of the river on the National Priorities List ("NPL"). The ensuing litigation over which entities were responsible for what share of the considerable investigation and cleanup costs has traced an eight-year oscillation through and between various levels of the federal court system. In the latest appearance in our courthouse, Plaintiff-Appellant Kalamazoo River Study Group ("KRSG"), a consortium of former papermill owners whose facilities polluted the river, appeals two distinct decisions of the District Court for the Western District of Michigan regarding the allocation of investigation and remediation costs to Defendants-Appellees Rockwell/Meritor ("Rockwell") and Eaton Corporation ("Eaton").

Following the placement of a stretch of the river on the NPL, the member companies of the KRSG entered into a remediation agreement with state and federal authorities. KRSG then filed actions against several other manufacturers who operated facilities on the river, including Rockwell and Eaton, under the contribution provision of the Comprehensive Environmental Response, Compensation and Liability Act of 1980 ("CERCLA"). See 42 U.S.C. § 9613(f). Following lengthy trials, punctuated by visits to our court, the district court found both Rockwell and Eaton liable for some of the PCB contamination, but allocated none of the investigation or cleanup costs to Rockwell and only a small portion to Eaton.

KRSG first appeals the district court's denial of its motion to reopen the order relieving Rockwell of any contribution responsibility. After the district court issued its zero-allocation order, KRSG discovered new evidence of increased environmental contamination, prompting it to file the motion to reopen. The district court, construing KRSG's filing as a motion under Rule 60(b) (2) of the Federal Rules of Civil Procedure, denied the request as time-barred. KRSG also appeals the district court's order allocating to Eaton only a small portion of the investigation costs and none of the future remediation costs. KRSG contends that the district court applied an inappropriate standard of liability and that the district court made several errors in its factual findings.

We AFFIRM both district court judgments.

In their nearly decade-long battle, these adversaries have amassed a prodigious factual record, brimming with environmental assessments, ecological data, and scientific opinions. We have already comprehensively detailed many of the pertinent factual disputes elsewhere. See Kalamazoo River Study Group v. Rockwell Int'l Corp., 171 F.3d 1065 (6th Cir. 1999) ("Rockwell I"); Kalamazoo River Study Group v. Menasha Corp., 228 F.3d 648 (6th Cir. 2000); Kalamazoo River Study Group v. Rockwell Int'l Corp., 274 F.3d 1043 (6th Cir. 2001) ("Rockwell II"). However, a brief overview of the litigation is in order.

In 1990, the federal Environmental Protection Agency ("EPA") added to the NPL a thirty-five-mile stretch of the Kalamazoo River after discovering, in coordination with the Michigan Department of Natural Resources ("MDNR"), large PCB concentrations in the river. Rockwell I, 171 F.3d at 1066. In conjunction with the EPA, the MDNR identified several paper mills owned by HM Holdings/Allied Paper, Inc. ("Allied"), Georgia-Pacific Corp. ("GP"), and Simpson Plainwell Paper Co. ("Simpson") as the main sources of the PCB contamination. In December 1990, these three mill owners entered into an Administrative Order by Consent ("AOC") with the MDNR, which required them to fund a remedial investigation of the NPL site, even though these KRSG members did not admit liability for the PCB pollution by signing the AOC. The Fort James Operating Co. ("Fort James"), which also owned a facility located adjacent to the NPL site, joined together with Allied, GP, and Simpson to form the Kalamazoo River Study Group, which would conduct the investigation and clean-up of the river. Rockwell I, 171 F.3d at 1067.

The study of the river, referred to as the Remedial Investigation/Feasability Study ("RI/FS"), uncovered massive PCB contamination. The AOC mandated that for the purposes of the RI/FS, KRSG had to study an expanded ninety-five-mile stretch of the Kalamazoo River. The RI/FS zone included the Eaton Battle Creek plant fifteen miles upstream of the NPL site and the Rockwell Universal Joint facility, located in Allegan downstream of the NPL site. The administrative successor to the MDNR, the Michigan Department of Environmental Equality ("MDEQ"), ultimately concluded in 1997 that the river contained over 350,000 pounds of PCBs.

PCBs, which accumulate predominantly in organically rich, quiescent areas of the river, present a grave public health risk, mainly because they contaminate fish with potentially cancerous chemical waste. Monsanto Corporation produced several different varieties of PCBs all under the brand name "Aroclor," (i.e., Aroclors 1242, 1254, and 1260, with the higher number corresponding to a greater PCB molecular weight). The KRSG companies used Aroclor 1242 extensively in their de-inking and paper manufacturing operations for several decades beginning in the 1930s, and they also used Aroclor 1254 in transformers, capacitors, hydraulic systems, and paints. Throughout the NPL site, Aroclor 1242 is the most prevalent PCB, and the MDEQ and EPA determined that KRSG companies were responsible for the bulk of the Aroclor 1242 contamination at the NPL site. The MDEQ also detected Aroclors 1254 and 1260 at the site, the questionable source of which forms the kernel of these appeals.

All four KRSG companies have not disputed that they are liable and responsible parties within the meaning of CERCLA. See 42 U.S.C. § 9607. They have, however, claimed that they only contributed minimally to the Aroclor 1254 and 1260 pollution at the NPL site and thus another party must be responsible for those PCBs. As responsible parties with a statutory right to contribution from potentially liable parties, see 42 U.S.C. § 9613(f), the KRSG members brought an action in December 1995 against Rockwell, Eaton, and six other companies, including Benteler Industries and Consumers Power. KRSG alleged that these factory owners were partially responsible for the PCB contamination at the NPL site such that they owed KRSG contribution for the costs of the investigation and future clean-up.

KRSG's contribution action against Rockwell alleged that Rockwell's Allegan facility, which produced universal joints for automobiles and construction equipment, contributed to the PCB contamination at the NPL site. The Allegan facility was in operation from the early 1900s until 1989. In 1987, the EPA, in an unrelated action, added the Allegan facility to the NPL because of arsenic, cyanide, and chromium contamination, but not because of any widespread PCB contamination. Soil tests at Allegan did reveal the presence of some Aroclor 1254 (along with some Aroclor 1242 and 1260) in the groundwater and light non-aqueous phase liquid (the oily film on top of water). Even though there existed no definitive proof that Rockwell ever purchased PCBs or conducted manufacturing operations that would necessitate the use of PCB-containing oils, the presence of PCBs on the property indicated that it did in fact purchase and employ PCBs in its operation. This fact, however, did not resolve the more salient question of whether these PCBs actually found their way to the Kalamazoo River such that they contributed to the overall pollution for which KRSG was responsible. In December 1998, the district court held that Rockwell did release PCBs into the Kalamazoo River NPL site and was therefore liable for some of the pollution.1 

Subsequently, in June 2000, the district court set out to determine the precise allocation of costs Rockwell owed to KRSG. The district court focused almost exclusively on the quantity of PCBs released by Rockwell versus the amount leaked by the KRSG companies. It determined that " [g]iven the low levels of PCBs on the Rockwell property, and the fact that the river sediments and the fish tend to show no significant contribution by Rockwell, the Court finds that Rockwell's PCB contribution was very minimal, particularly in contrast to the contribution by [KSRG's] members." Joint Appendix ("J.A.") I at 926 (Dist.Ct.Op.6/3/00).2  The court ruled that KRSG could not recover from Rockwell. KSRG appealed, arguing that the existence of a logical discontinuity between holding a party liable and allocating no costs demonstrated that the district court had abused its discretion. We affirmed the district court's holding in December 2001; there was no inconsistency, and the district court had broad discretion to allocate the costs of the remedial investigation, even if the result was a zero allocation to a liable corporation. See Rockwell II, 274 F.3d at 1049.

The immediate circumstances giving rise to this appeal were borne of the acrimonious relationship between the EPA and Rockwell. In April 1998, the EPA revoked Rockwell's authority to investigate the Allegan NPL site because of multiple delays and acts of noncompliance by Rockwell. Throughout 2000 and 2001, the EPA assumed control of the investigation and discovered that the Allegan facility in fact had dramatically higher PCB levels than Rockwell had previously disclosed; in some instances the new PCB levels were more than one hundred times the previously reported levels. On August 9, 2001, the EPA ordered Rockwell to abate an "imminent and substantial endangerment to the public health...." J.A. I at 942 (EPA Order, Aug. 2001). The EPA also noted one PCB plume was then entering the Kalamazoo River and another was migrating towards the river.

After learning of the EPA order, KRSG filed a motion with the district court on September 21, 2001 to reopen the CERCLA allocation proceedings fifteen months after the court had made its June 2000 allocation order. KRSG claimed that Rockwell had deliberately obfuscated this data in contravention of its statutory duty, and as a result KRSG asked the court to use its "equitable power" to reconsider the allocation. KRSG did not refer to its motion as a Rule 60(b) motion, but on October 15, 2001, the district court ruled as if the motion were of the Rule 60(b) (2) mold. The district court denied KRSG's motion because KRSG filed it after the one-year time limit for Rule 60(b) (2) motions had expired.

On appeal, KRSG offers several alternative arguments. First, KRSG contends that the district court erroneously considered KRSG's motion as a Rule 60(b) (2) motion, because CERCLA itself provides an equitable basis for reopening an allocation order in the face of changed circumstances. Second, KRSG argues that if Rule 60(b) does apply, then the district court should have invoked Rule 60(b) (5) instead of Rule 60(b) (2) because the district court's original order was "prospective" and accordingly subject to Rule 60(b) (5)'s "reasonable time" limitation, as opposed to Rule 60(b) (2)'s one-year time bar. Third, KRSG implicitly posits that the newly discovered evidence provides a sufficient basis for reallocating the remediation costs.

KRSG originally brought a contribution action against Eaton because Eaton operated three facilities near the Kalamazoo River NPL site: Eaton Marshall, Eaton Battle Creek, and Eaton Kalamazoo. The district court granted summary judgment in favor of Eaton as to the Marshall and Kalamazoo facilities in June 1998. Eaton Battle Creek, which is no longer in operation, was located fifteen miles upstream of Morrow Lake Dam and twenty miles upstream of the Kalamazoo River NPL site. While it was beyond the boundaries of the NPL site, it was within the larger confines of the RI/FS zone. The Battle Creek facility manufactured automotive parts, namely engine valves and gears, and undisputedly released significant quantities of oil into the river for over four decades. The main question attendant to the contribution action against Eaton is whether that oil contained PCBs and if it did, whether those PCBs affected the NPL site and the RI/FS zone.

There is some evidence that Eaton Battle Creek employed PCB-laden oils in its manufacturing processes, although it appears that it did not do so on a regular basis. Investigators discovered PCBs (primarily Aroclors 1248, 1254, and 1260) in several sewer outditches and in the wood blocks that lined the floor at Battle Creek. The district court eventually concluded that "the PCBs used in Eaton's Battle Creek facility were only found in the transformers and capacitors and the hydraulic fluids, and those fluids were not released to the River in any regular or measurable manner." J.A. II at 300 (Dist. Ct.Op.12/7/98). In other words, Eaton minimally used some PCB-containing hydraulic oils in closed systems; the oil from these systems did not flush directly into the river and leaked only in small amounts onto the floor and possibly into various sewer ditches on the Eaton property.

KRSG attempted to prove that Eaton PCBs actually entered the river and contributed to the pollution of the NPL site or the RI/FS zone by presenting undisputed evidence that Morrow Lake, situated between Battle Creek and the NPL site, is polluted with primarily Aroclor 1254 and minimally Aroclor 1260. KRSG could not have polluted Morrow Lake, because it lies upstream of the KRSG facilities, so the ensuing questions were (1) whether Eaton polluted Morrow Lake and (2) whether any of the Morrow Lake PCB contamination impacted the NPL site such that upstream polluters of Morrow Lake should be held responsible as contributing parties.3  Morrow Lake is contaminated with Aroclor 1254 (constituting ninety percent of the total PCB contamination), and while there is evidence that Battle Creek's wastewater discharge ditch is heavily contaminated with Aroclors 1254 and 1260 at levels comparable to or exceeding the Aroclor 1242 contamination at the actual NPL site, other companies, including Clark Equipment, used this ditch. Evidence also exists that between Eaton Battle Creek and Morrow Lake, the PCBs appeared in high concentrations in areas with low amounts of organic material, suggesting that when one controls for the presence of organic content (a procedure known as "carbon normalization"), the PCB concentration of Aroclors 1254 and 1260 near Battle Creek was among the highest in the Kalamazoo River area.

In December 1998, the district court ruled that while the evidence showed that Eaton did utilize some PCBs in its electrical and hydraulic equipment, any PCB leaks were minimal in volume and sporadic in occurrence, such that Eaton was not liable for contribution. However, in assessing liability, the district court employed an improper liability standard, which caused us to remand the case for reevaluation. See Kalamazoo River Study Group v. Menasha, 228 F.3d at 661. In May 2001, the district court, while maintaining its previous holding that there was only the most scant evidence of a measurable PCB discharge into the NPL site from Battle Creek, ruled that it was "constrained to find that Eaton is liable for some PCB releases ... to the Kalamazoo River." J.A. II at 399.

The district court then proceeded to receive evidence on the allocation of response costs between KRSG and Eaton. KRSG presented the testimonies of Brown and McLaughlin, with the latter stating that he discovered at Battle Creek PCB concentrations ranking in the upper 2% of all PCB samples taken from the NPL site. Many of McLaughlin's samples, which demonstrate the presence of high PCB concentrations, came from a drainage ditch leading to the river that Eaton shared with Clark Equipment. McLaughlin concluded that Battle Creek significantly contributed to the PCB contamination at the NPL site.

Eaton, in response, relied on the testimony of Connolly, who testified that the PCB concentrations in the ditch and several of the sample locations near the ditch's outfall to the river were not characteristic of the PCB concentrations in the river between Battle Creek and Morrow Lake. Connolly essentially posited that the Aroclor 1248, 1254, and 1260 concentrations in the general river channel were not consistent with Eaton Battle Creek as a PCB point source. Connolly noted the lack of a PCB gradient spanning the distance between Battle Creek and Morrow Lake. Generally, PCB concentrations will be highest near the PCB source and will increasingly diminish as the distance from that source grows. Comparing the Kalamazoo River to other PCB-contaminated rivers with a "normal" PCB gradient, such as the Hudson River in New York and the Housatonic River in Massachusetts, Connolly highlighted the lack of a gradient stretching from the Battle Creek "source" to further downstream points. Connolly testified that this lack of a gradient contradicted the conclusion that Battle Creek was a primary contributor of Aroclor 1254. Connolly bolstered his theory by suggesting that the absence of a high concentration of PCB contamination at the sediment-rich upstream entrance to Morrow Lake provided strong evidence that facilities closer to Morrow Lake, such as the Benteler Industries and Consumers Power sites, were the true sources of the Morrow Lake contamination.

On August 29, 2002, the district court ordered Eaton to pay ten percent of KRSG's investigation costs in the RI/FS zone (ten percent of the total $622,615.79 investigation cost for that portion equals $ 62,261.58). First, the court noted that KRSG had "the burden of proving its equitable right to contribution by a preponderance of the evidence." J.A. II at 433 (citing United States v. R.W. Meyer, Inc., 932 F.2d 568, 573-74 (6th Cir. 1991)). Second, the district court rejected McLaughlin's testimony in favor of Connolly's, writing that while it did "not find Mr. McLaughlin's conclusions to be well supported," it "found the testimony of Dr. Connolly to be more persuasive...." J.A. II at 443, 445. It reached this conclusion because of Connolly's credentials and because it had more confidence in the science underlying Connolly's conclusions. In listing the reasons why it disfavored McLaughlin's position, the district court noted that KRSG's claim that Eaton principally contaminated the river with Aroclor 1254 "ignore [d] other potential sources of PCBs located upstream of Eaton's Battle Creek facility that could have contributed to the PCBs found in the former channel and current channel of the Kalamazoo River near the Eaton/Clark ditch." J.A. II at 444. Consequently, the court agreed with Connolly's dual contentions that other parties contributed to the PCB contamination in Morrow Lake and that Eaton discharged only a "de micromis," J.A. II at 457, amount of PCBs to the Eaton/Clark ditch. The court found persuasive the argument that only a fraction of this "de micromis" amount traveled to Morrow Lake and only a fraction of that fraction actually passed over the Morrow Lake Dam to the NPL site.

The court then reached the ten-percent figure by applying the "Gore Factors" outlined in Centerior Service Co. v. Acme Scrap Iron & Metal Corp., 153 F.3d 344, 354 (6th Cir. 1998), and relying almost exclusively on the "the amount of the hazardous waste involved." Plaintiff's experts had estimated that Eaton was responsible for forty percent of the Aroclors 1254 and 1260 at the NPL site, but the district court rejected this contention as lacking an "articulated scientific basis." J.A. II at 468. Instead, the district court accepted Connolly's estimate that only 1.3% of the PCBs in the NPL site could have come from Morrow Lake. In sum, the court rejected both the notion that Morrow Lake was a primary contributor of PCBs to the NPL and that Eaton was a primary contributor to the PCB contamination at Morrow Lake.

On appeal of the Eaton decision, KRSG first contends that " [a]lthough the District Court paid lip service to the preponderance of the evidence standard ... it applied a much more stringent standard," requiring KRSG to disprove the potential responsibility of other upstream facilities. KRSG's Eaton Br. at 17. Second, KRSG claims that the district court committed clear factual errors by accepting the theory that the PCB contamination at Morrow Lake and at the NPL site might have come from sources other than Battle Creek.

The district court had proper jurisdiction over the original contribution action pursuant to 42 U.S.C. § 9613(b), the jurisdictional provision for CERCLA actions. The district court's ability to hear KRSG's motion to reopen the allocation proceedings stems from those same jurisdictional provisions. In a November 9, 2001 order, we ruled that we had jurisdiction over KRSG's appeal from the district court's denial of the motion to reopen the allocation proceedings against Rockwell. We have jurisdiction over KRSG's appeal of the district court's final Eaton allocation decision pursuant to 28 U.S.C. § 1291.

Because KRSG's two appeals present completely distinct legal and factual issues, we discuss each in turn. First, we examine KRSG's claim that the district court erred in denying KRSG's motion to reopen the allocation decision against Rockwell. Second, we explore KRSG's argument that the district court erred both in its choice of a standard of liability and its application of that standard to the facts in its Eaton allocation order. We hold that the district court acted properly in both instances.

Generally, we review the denial of a Rule 60(b) motion for abuse of discretion. See Blue Diamond Coal Co. v. Trustees of the UMWA Combined Benefit Fund, 249 F.3d 519, 524 (6th Cir. 2001). However, we must "treat the district court's interpretation and application of the Federal Rules of Civil Procedure as a question of law and, as with all legal questions, review [the district court's] analysis de novo." Jalapeno Prop. Mgmt., LLC v. Dukas, 265 F.3d 506, 510 (6th Cir. 2001) (citing Indiana Lumbermens Mut. Ins. Co. v. Timberland Pallet & Lumber Co., 195 F.3d 368, 374 (8th Cir. 1999)). The district court's decision to construe KRSG's motion to reopen as one falling under the umbrella of Rule 60(b), and more specifically Rule 60(b) (2), is clearly an interpretation and application of the Rules of Civil Procedure, and thus we review de novo this portion of the court's analysis.

2. The Reopening of Allocation Orders under CERCLA

KRSG bases its objection to the district court's pigeonholing of its motion to reopen as a Rule 60(b) motion on the notion that the inherently equitable nature of the CERCLA allocation mechanism permits the reopening of allocation judgments independent of Rule 60(b). KRSG argues that CERCLA allocation orders are subject to revision whenever the equities underlying the decision shift. Finding no aspect of CERCLA that confirms KRSG's assertion, we disagree.

KRSG is certainly correct that principles of equity guide CERCLA's contribution provision, but nothing in CERCLA compels the conclusion that the equitable underpinnings of an allocation decision exempt it from the requirement that motions to alter judgments be brought under Rule 60(b). CERCLA permits courts to "allocate response costs among liable parties using such equitable factors as the court determines are appropriate." 42 U.S.C. § 9613(f) (1). There is nothing in § 9613(f) that suggests that Rule 60(b) is not the proper vehicle for altering a judgment, and quite the opposite, 42 U.S.C. § 9613(f) (1) clearly states that all claims "shall be brought in accordance with ... the Federal Rules of Civil Procedure." Presumably, this illustrates Congress's intention to make CERCLA allocation decisions no less subject to the Federal Rules of Civil Procedure than any other contribution decision.

The crux of KRSG's argument is that because a district court relies upon equitable factors to make an allocation decision, such a decision is forever subject to revision should there be any alteration in the equities underlying the allocation order. KRSG's position cannot prevail. The equitable basis of CERCLA allocation decisions does not deprive all allocation orders of their finality. Other equitable decisions, such as an order mandating specific performance in a contract dispute, are not automatically subject to future revision. KRSG does not point us to any part of CERCLA in which Congress has expressed a desire that all allocation decisions should be considered ongoing or nonfinal such that there is another method by which relief from judgment may be sought other than Rule 60(b).4 

KRSG directs our attention to two cases as support for its view that CERCLA contains an internal mechanism for reopening allocation decisions, but neither opinion persuasively proves KRSG's contention. In Acushnet Co. v. Coaters, Inc., 972 F. Supp. 41(D. Mass. 1997), a federal district court fashioned an ongoing and provisional CERCLA allocation order, stating, "Either party may file with the court, when good cause to do so has developed factually, a motion supported by a showing of a material change in circumstances that justifies a change in the allocation of shares among the parties." Id. at 63. Uncertainty about the completeness of the remedial-cost evidence before the court prompted it to permit explicity a "reasonable opportunity for any interested party to initiate later proceedings to modify the provisional allocation of equitable shares of legal responsibility...." Id. at 69. The district court believed that aspects of CERCLA militated against finality, and it accepted "the consequences of delay and greater expense of final adjudication in order to come closer... [to] equitable allocation of legal responsibilities." Id. at 62. The court issued a provisional judgment because of the insufficiency of the evidence before it, which prevented it from assessing "equitable shares of legal responsibility with the degree of confidence implicit in findings made on a preponderance of the evidence." Id. at 71.

KRSG also cites the Seventh Circuit's opinion in PMC, Inc. v. Sherwin-Williams Co., 151 F.3d 610 (7th Cir. 1998), for the proposition that district courts may always reopen an allocation judgment even in the absence of a Rule 60(b) motion. In PMC, the district court had made Sherwin Williams liable for one-hundred percent of the costs of cleaning up a hazardous waste site. Sherwin Williams sought contribution because PMC had dumped waste at the site after it acquired the property from Sherwin Williams. The court recognized that PMC might be responsible for future clean-up costs should it be ordered to clean up any waste, other than waste at issue in Sherwin Williams's contribution order, that PMC contributed to the site after it purchased the property. Id. at 617. However, the Seventh Circuit held that the district judge did not abuse his discretion in ruling that PMC did not owe Sherwin Williams contribution for already incurred costs because PMC's dumped waste "may have been too inconsequential to affect [Sherwin Williams'] cost of cleaning up significantly." Id. at 616. KRSG directs our attention to the Seventh Circuit's consideration of whether the allocation of cleanup costs was premature given that it concerned costs "that PMC has not yet incurred." Hypothesizing, the court confirmed that the allocation was proper because " [i]t economizes on judicial time ... and it also lets the parties know at the earliest opportunity where they stand." Id. The court continued, "cooperativeness in doing the actual clean-up is a relevant equitable factor that cannot be evaluated until the clean up is complete ... [b]ut this concern can be accommodated ... by allowing the district court to make an all-at-once determination subject to the court's revisiting the issue should a failure of cooperation or some other unforeseen circumstance make adherence to the original determination inequitable." Id. (citations omitted).

What these cases show is not that allocation decisions in CERCLA cases are inherently subject to change, but rather that courts have the power to fashion relief that is subject to future change. Neither case stands for the proposition that CERCLA provides an alternative route for reopening decisions in lieu of Rule 60(b), but rather they affirm the broad equitable powers of the district court. The court in Acushnet saturated nearly thirty written pages with an explanation for why a provisional ruling was necessary. While the Acushnet court did discuss an ideal of flexibility within CERCLA, it did not (nor could it) establish a broad rule that allocation orders were provisional and exempt from Rule 60(b). It instead created a provisional order in the face of insufficient evidence when there was a concern that the evidentiary moorings of any fixed allocation would disintegrate in the future.

Similarly, KRSG's focus on PMC is misplaced. The Seventh Circuit's PMC decision only hypothesized about what a district court could do in the face of a premature claim if there were uncertainty about one of the equitable factors, i.e., cooperation. The Seventh Circuit's statements in dicta did not establish a ground rule that all allocation decisions based on equitable determinations will always be subject to revisions. Moreover, hurting rather than helping KRSG's argument is the fact that the Seventh Circuit affirmed the district court's decision not to alter the allocation despite new evidence of contamination when the district court concluded that any contribution PMC may have made to the contamination at the site was negligible.

In allocating no costs of the future remediation to Rockwell, the district judge mentioned nothing about a provisional order or potential alterations in the future. Nor does the language of the district court's order leave room for us to infer that the allocation decision was provisional or susceptible to change based upon future events. Unlike the lower courts in Acushnet and PMC, the district court did not describe how the circumstances of this case or the insufficiency of the evidence left open the possibility for future alteration of the allocation. Rather, the district court here simply stated that under the "Gore" equitable factors, "Rockwell should not be required to contribute to the remediation of the ... Superfund site. The PCB releases by Plaintiff's members are more than sufficient to justify imposing on Plaintiff the entire cost of response activities relating to the NPL site." J.A. I at 926 (Dist.Ct.Op.6/3/00). This allocation order does not intimate that the evidence before the district court on any of the equitable factors considered was incomplete such that the order would be subject to revision without a Rule 60(b) motion. Nothing in our opinion, however, should be construed as ruling that district courts cannot fashion provisional allocation orders or that district courts must explicitly label an allocation order "provisional" for a reviewing court to evaluate it as such. As demonstrated by Acushnet and PMC, provisional allocation orders can be valuable equitable tools in the event of incomplete evidence.

This was not such an order. There is no reading of the district court's opinion that suggests its zero-allocation order was ongoing or subject to change in the future. Furthermore, we have been presented with no support for the notion that CERCLA provides a mechanism independent of Rule 60(b) for revisiting allocation orders.

3. The District Court's Choice Between Application of Rule 60(b) (2) or Rule 60(b) (5)

KRSG alternatively contends that if its motion to reopen is viewed as a Rule 60(b) motion, then the more generous time limits of Rule 60(b) (5) should apply. The district court analyzed KRSG's motion as if it were a Rule 60(b) (2) motion and accordingly denied it as time-barred, because KRSG filed its motion more than one year after the entry of the allocation order.5  We review de novo the district court's decision and hold that there was no error, because the district court's initial allocation order was not "prospective" within the meaning of Rule 60(b) (5) and accordingly the time strictures of Rule 60(b) (2) control.

Rule 60(b) (5) states in pertinent part that "the court may relieve a party... from a [n] ... order ... [when] it is no longer equitable that the judgment should have prospective application." Unlike Rule 60(b) (1)-(3) motions, which cannot be brought more than a year after an entry of judgment, Rule 60(b) (5) motions can "be made within a reasonable time" after the entry of the order. Fed. R. Civ. P. 60(b) (5). The application of Rule 60(b) (5) here turns on the meaning of "prospective." The mere possibility that a judgment has some future effect does not mean that it is "prospective" because " [v]irtually every court order causes at least some reverberations into the future, and has ... some prospective effect." Twelve John Does v. District of Columbia, 841 F.2d 1133, 1138 (D.C. Cir. 1988). The essential inquiry into the prospective nature of a judgment revolves around "whether it is `executory' or involves the `supervision of changing conduct or conditions.'" Id. at 1139 (quotation omitted). Injunctions (permanent or temporary), some declaratory judgments, and particularly consent decrees are prospective judgments susceptible to a Rule 60(b) (5) challenge. See 12 James Wm. Moore, Moore's Federal Practice §§ 60.47 [1]- [2]. Money judgments, however, do not generally have prospective application because they are final in the sense of involving a set monetary outlay. Id. at § 60.47 [1] [b].

Most cases consider Rule 60(b) (5)'s "prospective application" clause in the context of consent decrees, which are prospective by nature. The Supreme Court in Rufo v. Inmates of Suffolk County Jail, 502 U.S. 367, 112 S. Ct. 748, 116 L. Ed. 2d 867 (1992), held that a plaintiff in institutional-reform litigation could get relief under Rule 60(b) (5) where "a significant change either in factual conditions or in law" altered the equitable basis for an ongoing consent decree. Id. at 384, 112 S. Ct. 748. The Court made clear in Rufo that this rule should not be limited to institutional-reform litigation, but while the Court may have intended its analysis to apply to consent decrees involving private parties, the Court did not expand the scope of the term "prospective" so as to encompass more varieties of equitable judgments. We are not aware of any case in which Rufo has been applied to judgments other than consent decrees, declaratory judgments, and injunctions, which often require ongoing court supervision and future judicial involvement. See Lorain NAACP v. Lorain Bd. of Educ., 979 F.2d 1141, 1148 (6th Cir. 1992) (citing the Rufo modification standard in the school desegregation context); see also United States v. W. Elec. Co., 46 F.3d 1198, 1203 (D.C. Cir. 1995) (applying Rufo's "less stringent [ ] standard" to antitrust consent decree outside of the institutional-reform context, but noting that Rufo only applies to certain types of injunctive relief) (alteration in original); United States v. Kayser-Roth Corp., 272 F.3d 89, 95-96 (1st Cir. 2001) (engaging in a Rule 60(b) (5) analysis when a declaratory judgment for future response costs would have prospective effect and there existed changed circumstances).

We also must heed the requirement that parties cannot disguise Rule 60(b) (1)-(3) motions as 60(b) (4)-(6) motions. What in reality is a 60(b) (2) motion cannot be labeled as a 60(b) (5) motion to gain the benefits of a more generous limitations period. McDowell v. Dynamics Corp. of Am., 931 F.2d 380, 383-84 (6th Cir. 1991). KRSG complains that the new evidence of PCB dumping at the Allegan facility alters the equitable calculus employed by the district court, but its claim sounds very much like a claim regarding newly discovered evidence, which is controlled by Rule 60(b) (2). See CMC Heartland Partners v. Union Pac. R.R. (In re Chicago, Milwaukee, St. Paul & Pac. RR Co.), 78 F.3d 285, 293-94 (7th Cir. 1996) (considering under Rule 60(b) (2) claim that party was entitled to relief when there was new evidence regarding cleanup of site in an action concerning a noncontribution provision of CERCLA).

KRSG is incorrect in its assertion that the district court's allocation order was "prospective" in the Rule 60(b) (5) sense of the word. The district court's allocation order was not a consent decree, an injunction, or even a declaratory judgment. Rather, the allocation decision stated that Rockwell was not responsible for any measurable PCB contamination to the NPL site; this was a one-time judgment that Rockwell was not required to contribute and it did not provide for any future supervision or alteration by the district court. Merely because KRSG requested contribution for future costs, which the district court denied, and merely because KRSG's prospective remediation expenses would be higher in a relative sense as a result of the district judge's order, does not mean that the order was "prospective" under Rule 60(b) (5).

Thus, we agree with the district court's decision to apply Rule 60(b) (2), which permits motions for relief from orders based upon "newly discovered evidence which by due diligence could not have been discovered in time to move for a new trial under Rule 59(b)." Fed. R. Civ. P. 60(b) (2). Rule 60(b) (2) motions must be brought within one year after entry, which permits a party an extra 355 days more than the usual ten-day period to file a motion for a new trial. KRSG's motion must fail because it was filed more than fifteen months after the entry of the judgment.6 

KRSG appeals both the evidentiary standard employed by the district court and the court's allocation to Eaton of only for ten percent of the investigation and for none of the future cleanup costs. More specifically, KRSG claims that the district court paid "lip service" to the "preponderance of evidence" standard, while actually employing a much higher and more stringent standard; in KRSG's estimation, the district court held it to an impermissible standard of "absolute disproof" because it required KRSG to demonstrate conclusively that other parties did not cause the PCB contamination in the discharge ditch and in the river. We disagree. We conclude that the district court applied the appropriate "preponderance of the evidence" standard and neither abused its discretion in allocating only ten percent of the costs to Eaton nor committed a clear error in its factual findings.

We review different aspects of the district court's order using different standards of review. We review de novo the legal conclusions of the district court, but we review the court's factual findings following a bench trial for clear error. Menasha, 228 F.3d at 652. We review the district court's allocation decision more deferentially, particularly because of the equitable nature of CERCLA contribution decisions. See Franklin County Convention Facilities Auth. v. Am. Premier Underwriters, Inc., 240 F.3d 534, 549 (6th Cir. 2001) (noting the district court's "broad discretion" in allocating CERCLA contribution). We will overturn the district court's allocation order only if we are left with the "definite and firm conviction that the trial court committed a clear error of judgment." Rockwell II, 274 F.3d at 1047 (quoting Logan v. Dayton Hudson Corp., 865 F.2d 789, 790 (6th Cir. 1989)). The factual findings underlying the district court's allocation order are reviewed for clear error. Schroyer v. Frankel, 197 F.3d 1170, 1173 (6th Cir. 1999); see also Franklin County, 240 F.3d at 541 (noting in a CERCLA contribution opinion that " [w]here two logically permissive interpretations of the evidence exist, the trial judge's selection cannot be adjudged clearly erroneous on appeal.")

2. The District Court's Application of a Standard of Liability

The district court's application of a preponderance of the evidence standard is a mixed question of law and fact that we review de novo. CERCLA contribution plaintiffs must prove that they are entitled to reimbursement by a preponderance of the evidence. See B.F. Goodrich v. Betkoski, 99 F.3d 505, 526 (2d Cir. 1996). Nothing in the wording of the CERLCA contribution provision suggests otherwise. See 42 U.S.C. § 9613(f). In his August 2002 allocation order, the district judge wrote that "although [KRSG] is not required to prove its case with ... mathematical precision, or scientific certainty, it still has the burden of proving its equitable right to contribution by a preponderance of the evidence." J.A. II at 433 (Dist.Ct. Op.8/29/02). KRSG asserts that the district court only nominally applied this standard, while effectively holding KRSG to a much higher standard. This argument is misguided because the district judge specifically stated that he was employing a preponderance of the evidence standard and even made clear that KRSG did not have to prove its case with "scientific certainty," or "absolute proof" as KRSG labels it. KRSG's argument that the allocation order focused too heavily on KRSG's inability to disprove the possible contamination by other facilities on the Kalamazoo River is a contention that the district court misweighed the evidence; it is not an assertion that the district court applied an inappropriate burden of proof.

We affirm the district court's judgment because the court neither clearly erred in its factual finding nor abused its discretion in allocating only a small portion of KRSG's investigation costs to Eaton. KRSG contends that the district court improperly required KRSG to disprove the complicity of potentially responsible parties other than Eaton. Eaton responds that KRSG did not provide enough evidence that Eaton could ever be "ruled in" as a PCB contributor and thus that KRSG's failure to "rule out" other potential contributors was only one part of the district court's final order. We accept neither view, but rather affirm the district court's order because it appropriately weighed the evidence presented by both sides to conclude that Eaton was only minimally responsible for some of the investigation costs incurred by KRSG.

First, there is no indication that the district court clearly erred in its factual findings. Some of the evidence does support the theory that Eaton contributed to the PCB contamination for which KRSG is responsible. Aroclors 1254 and 1260, which comprise twenty-five percent of the PCB contamination at the NPL site and ninety percent of the PCB contamination at Morrow Lake, were found at Battle Creek. Throughout its history, Eaton Battle Creek discharged large quantities of oil to the river, and the wastewater ditch connecting the Battle Creek facility to the river contained high PCB concentrations, particularly when one controls for the amount of carbon content. Additionally, KRSG's expert, McLaughlin, testified that no other potentially responsible party upstream of Morrow Lake, including Clark Equipment with whom Eaton shared the wastewater ditch, released as much PCBs to the river as Eaton. McLaughlin also concluded that sources closer to Morrow Lake, such as Benteler Industries and Consumers Power, did not release PCBs to Morrow Lake and thus Eaton must be responsible.

Eaton countered KRSG's evidence with support of its own and ultimately the court found Eaton's proof more compelling than KRSG's. The evidence presented by Eaton showed that it barely used any PCB-laden oils at Battle Creek and that most of the PCBs found in the wood floors came from closed — system transformers and hydraulics. The PCBs in the floor were thus the result of incidental and minor spillage, the oil from which would not have been flushed to the river. Eaton's expert explained that polluted rivers generally have a higher concentration of PCBs near the source with decreasing concentrations as one travels further from the source. There was no such gradient stretching downstream from Battle Creek, and the MDNR found scant proof of PCB contamination in the fifteen-mile stretch of river between Battle Creek and Morrow Lake, which indicated that sources closer to Morrow Lake, such as Benteler Industries and Consumers Power, primarily contaminated the lake with PCBs. Such a conclusion was bolstered by the fact that the downstream portions of Morrow Lake had higher PCB concentrations than the entrance to the lake closer to Battle Creek; if there was an upstream PCB source, the upstream entrance to Morrow Lake should have had high PCB levels because of its high organic content, but in reality it had a low PCB concentration.

The district court assessed all of this evidence and concluded that Eaton's use of PCBs was exceedingly minimal and that any PCBs it did use barely impacted the pollution at Morrow Lake, let alone the contamination at the actual NPL site. In making this factual finding, the district judge rejected the testimony of KRSG's expert McLaughlin because the judge did not believe that it was well-supported and because he considered Connolly's testimony to be more persuasive. This preference of one expert over another was not clearly erroneous, as the district judge had the opportunity to assess and observe each expert before making his decision. The district court noted that Connolly "has the most expertise of any of the witnesses in the area of hydrogeology and the transport of PCBs in the river environment" and that Connolly had testified before Congress. J.A. II at 445 (Dist.Ct.Op.8/29/02). Based on his acceptance of Connolly's testimony, the district judge concluded that KRSG had not proven by a preponderance of the evidence that Eaton should be responsible for a portion of KRSG's costs. The district court did not clearly err in making these factual determinations; they were well-supported by the evidence as a whole, and the district court chose between two permissible interpretations of the evidence.

KRSG contends that the district court erred because rather than requiring it to prove Eaton's complicity, the court forced KRSG to disprove the potential responsibility of Clark, Benteler, and Consumers Power to the PCB contamination in the Eaton/Clark wastewater ditch and Morrow Lake. This argument is flawed. Most importantly, the court did not rely solely on evidence that other parties may have added to the PCB contamination in Morrow Lake and the NPL site; in making its ruling, the court focused primarily on Connolly's testimony about the lack of a gradient and the evidence that any PCBs Eaton may have leaked did not contribute to the pollution upstream of Morrow Lake, in Morrow Lake, or in the NPL site. Additionally, evidence exists that Clark, Benteler, and Consumers may have contributed to the PCB contamination upstream of the NPL site. The district court did not require KRSG to disprove in any absolute sense the potential contamination by those facilities, but rather considered the significant probability that they added to the pollution for which Eaton was being blamed. Thus, we cannot agree that the district court clearly erred in reaching its factual findings.

Second, we hold that the district court did not abuse its discretion by concluding that Eaton was responsible only for ten percent of the investigation costs and not responsible for any future costs of remediating the NPL site. KRSG contends that the district judge abused his discretion, because he ignored evidence of Eaton's culpability and required KRSG to disprove affirmatively that other parties, such as Clark and Benteler, were the actual sources of the PCBs. Again, we cannot accept KRSG's position. The district judge did not ignore evidence of Eaton's role in adding to the contamination, but rather found other contrary evidence more probative and persuasive. The district court noted that Eaton may have released some PCBs into the river, although it could not say with certainty that Eaton was the source of the PCBs in the wastewater ditch. Yet, the district court rested its decision on Connolly's testimony and ruled that it was unlikely that Eaton's PCBs added to the pollution in the Morrow Lake Dam and even less likely that these PCBs then entered the NPL site. If Eaton did release any PCBs, it released an extremely minimal volume to the RI/FS zone such that Eaton was responsible for only a small share of KRSG's costs.

Furthermore, the court did not myopically base its decision on KRSG's failure to rule out Clark and Benteler as potential PCB sources. Whereas KRSG views the district court as requiring it to disprove other parties' potential responsibility, the district court in reality recognized that some convincing evidence demonstrated that parties other than Eaton may have been the chief polluters of the wastewater ditch, Morrow Lake, and the NPL site. It accordingly held that Eaton was responsible for only a small portion of KRSG's costs, because evidence showed that Clark contributed to the PCB contamination in the discharge ditch and that Benteler Industries and/or Consumers Power were the main sources of Aroclors 1254 and 1260 at Morrow Lake.7  Yet, even ignoring the evidence of PCB contamination by Clark, Benteler, and Consumers Power, we are not left with the firm and definite conviction that the district court committed a mistake. Connolly's testimony about the lack of a gradient flowing downstream from Eaton and the concentration of PCBs at the downstream, rather than the upstream border of Morrow Lake, as well as the dearth of evidence that Eaton even used PCBs in significant quantities at any point in its operation, all help bolster the district court's allocation order.

Because the district court correctly dismissed KRSG's motion to reopen as a Rule 60(b) (2) motion that was time-barred and because the district court did not commit clear error or abuse its discretion in allocating to Eaton only a very small portion of KRSG's investigation costs, we AFFIRM the decisions of the lower court.

 *

The Honorable Richard Mills, United States District Judge for the Central District of Illinois, sitting by designation

 1

Although the district court improperly burdened KRSG with a higher liability standard than was appropriate and this panel subsequently reversed the district court's application of that standard, the district court's finding of liability for Rockwell stands under both the discredited higher standard and the less onerous standard See Kalamazoo River Study Group v. Menasha Corp., 228 F.3d 648, 655 (6th Cir. 2000).

 2

Owing to the consolidated nature of this case, the parties submitted two different joint appendices. The KRSG-Rockwell joint appendix (No. 01-2453) is denoted J.A. I, whereas the KRSG-Eaton joint appendix (No. 02-2192) is denoted J.A. II

 3

KRSG is not responsible for the clean-up of Morrow Lake and other areas outside of the defined NPL site, but the RI/FS zone encompasses Morrow Lake and the Kalamazoo River upstream to Eaton Battle Creek. Eaton, as a contribution defendant, is potentially liable for future remediation costs only to the extent that PCBs for which it is responsible flowed downstream and impacted the NPL site. Eaton could be responsible for the costs of investigating the RI/FS even if its PCBs did not affect the NPL site

 4

At oral argument, counsel for KRSG suggested that two sections of CERCLA support KRSG's argument in this regard See 42 U.S.C. §§ 9605(c) (4), 9622. Counsel's reference to these sections was misguided because while both concern the impact of future events upon some aspect of CERCLA, neither even remotely supports KRSG's argument. Section 9605(c) (4) is completely inapposite here; it concerns future changes to the President's national contingency plan for the removal of hazardous substances and states that " [n]othing ... shall preclude the President from taking new information into account in undertaking response actions...." Section 9622 is slightly more on point, as it controls the President's ability to enter into settlements and agreements with potentially responsible parties. It explicitly states that cleanup agreements are to be entered as consent decrees in a district court. 42 U.S.C. § 9622(d) (1) (A). Reliance on § 9622 does not aid KRSG. First, a consent decree is inherently different from a contribution allocation decision, which is a monetary judgment; a consent decree is "a settlement agreement subject to continued judicial policing." Lorain NAACP v. Lorain Bd. of Educ., 979 F.2d 1141, 1148 (6th Cir. 1992) (quotation omitted). Second, § 9622 does not explicitly discuss any mechanism for reopening final judgments that is unique to CERCLA Instead it states, "In the case of consent decrees ... no provision of this chapter shall be construed to preclude or otherwise affect the applicability of general principles of law regarding the setting aside or modification of consent decrees or other settlements." 42 U.S.C. § 9622(m).

 5

At the outset, we cannot criticize the district court's turn to Rule 60(b) (2), given that it was analyzing a motion to reopen that did not even mention Rule 60(b). The district court in many ways gave KRSG the benefit of the doubt, because the district judge could have denied the motion on the ground that his allocation order was not subject to revision absent a Rule 60(b) motion given that nothing in CERCLA mandates the reopening of allocation orders. However, the district judge did not follow that path, and now, after KRSG's failure to mention Rule 60(b) initially, KRSG contends that the district judge chose the wrong subsection to apply

KRSG's approach raises the specter of waiver. It is "a prerequisite to relief under Rule 60(b), [that] a party must establish that the facts of its case are within one of the enumerated reasons contained in Rule 60(b)...." Lewis v. Alexander, 987 F.2d 392, 396 (6th Cir. 1993). Failure to raise an issue at the trial court generally prevents a party from arguing the issue on appeal. See Thurman v. Yellow Freight Sys. Inc., 90 F.3d 1160, 1172 (6th Cir. 1996). KRSG claims that it did raise the substance of its Rule 60(b) (5) argument in its motion to reopen, even though it did not explicitly mention Rule 60(b) (5). As support, it points to a 1998 decision in which we stated that an argument not explicitly raised at the trial court level would still be considered when it involved a question of law. United Food & Commercial Workers Union, Local 1099 v. Southwest Ohio Reg'l Transit Auth., 163 F.3d 341, 360 n. 9 (6th Cir. 1998). While United Food is somewhat inapposite, because it concerned a party's attempt to raise a First Amendment overbreadth argument that was not raised in those terms initially, we agree that KRSG has eluded waiver by the slimmest of margins. The general principle that a district court should not raise a Rule 60(b) motion sua sponte does not apply, because the rule against doing so only adheres if the litigant has failed to file any motion, whereas here KRSG filed a motion, albeit an ambiguous one. See Eaton v. Jamrog, 984 F.2d 760, 762 (6th Cir. 1993).

 6

Even if KRSG had filed a timely Rule 60(b) (2) motion, it could not obtain relief. It did not appear to exercise due diligence in obtaining the new evidence. It is somewhat surprising that KRSG did not seek to test the Rockwell site itself sometime during discovery. It is even more incredible that KRSG did not seek access to the Rockwell site in April 1998 when the EPA announced its suspicion that Rockwell was "not fulfilling its obligation to thoroughly and objectively investigate the Site and present the results on behalf of the public; but, rather, contriving a study to avoid liabilities at the Site [,]" which raised a bold red flag. J.A. I at 1498 (EPA Letter). KRSG certainly had time to investigate the site either during the three years of discovery that preceded the liability portion of the trial or during the eighteen-month span between the April 1998 EPA action and the beginning of the allocation phase in November 1999

 7

KRSG's argument that the district court is being inconsistent with its past opinions in regards to the potential contributions of Clark and Benteler is without merit. KRSG argues at one point that the district court had previously exculpated Clark by stating that there was no qualitative or quantitative evidence of any PCB discharge by Clark See KRSG's Eaton Br. at 21 (citing J.A. II at 437 (Dist.Ct. Op.8/29/02)). Quite the opposite, the district court noted that Clark may have been responsible for some PCB contamination. It first remarked that even KRSG's expert, McLaughlin, "conceded that the PCBs found in the Eaton/Clark ditch could have originated from Clark if Clark's effluent contained PCBs." J.A. II at 437 (Dist.Ct.Op.8/29/02). The district judge cautioned that "no direct evidence" showed that Clark's discharges contained PCBs, but he mentioned that Clark may have used PCB-containing oils in its transformers and hydraulic systems. J.A. II at 437. Furthermore, there is no inconsistency with the district court's previous finding during the liability phase, see J.A. II at 382-83 (Dist.Ct. Op.5/9/01), that Clark did not discharge any effluent to the ditch it shared with Eaton, because the district court held that this earlier finding was based upon the erroneous testimony of Thomas Mattson, who mistakenly claimed that Clark did not discharge any PCB-containing effluent to the shared ditch. See J.A. II at 436-37 (Dist.Ct.Op.8/29/02).

KRSG also complains about an alleged inconsistency between previous findings and the district court's holding that facilities closer to Morrow Lake were in fact the most likely source of the Aroclor 1254 and 1260 pollution at Morrow Lake. In an earlier order, the district court granted Benteler's motion for summary judgment on KRSG's contribution claim, which we affirmed. Rockwell I, 171 F.3d at 1073. KRSG believes that the district court cannot have logically "exonerated" Benteler in 1996 and then found it to be a potential PCB contributor in 2002. KRSG's argument highlights its misunderstanding of the nature of its evidentiary burden. The district court did not exonerate Benteler. Rather, it granted the motion for summary judgment because KRSG did not offer evidence that created a genuine issue of material fact regarding Benteler's contribution of PCBs. Here, a different defendant, Eaton, presents the type of evidence that KRSG should have presented six years ago; i.e., some convincing facts that the Morrow Lake contamination did not come from sources upstream of Morrow Lake, but rather came from facilities on the lake. This is not an inconsistency, but rather it is the reality that Eaton is using evidence against KRSG that would have helped KRSG in its case against Benteler.