Full opinion text
MEMORANDUM — DECISION & ORDER McAVOY, District Judge. I. Procedural History On June 10, 1987, the United States of America and the State of New York (collectively “the government”) initiated an action against 83 business entities to recover response costs pursuant to Section 107 of the Comprehensive Environmental Response, Compensation and Liability Act (“CERCLA”) as amended, 42 U.S.C. § 9601, et seq., in connection with the clean-up costs of a hazardous waste site formerly owned by Pollution Abatement Services of Oswego, Inc. (“PAS”). Shortly after commencement of that action, the government entered into a consent decree with 82 of the defendants, recovering approximately $9.1 million of the approximately $12.3 million in response costs incurred through April 1,1987. The present litigation proceeded against Acan Aumi-num Corporation (“Acan”), the only non-settling defendant. In February 1988, Acan commenced a third-party action against Cornell University (“Cornell”). That action sought a declaration that Cornell was jointly and severally liable for the response costs incurred by the government relative to PAS and an order directing Cornell to contribute its fair share of the costs. In January of 1991, this Court granted summary judgment in favor of the government, holding Alcan jointly and severally liable for approximately $4 million in response costs at PAS. See United States v. Alcan, 755 F.Supp. 531 (N.D.N.Y.1991) (“Alcan-PAS”). The Court also granted Alcan’s motion against third-party defendant, Cornell, and found that a hearing was necessary to determine Cornell’s fair share of response costs. See id. On November 19 and 20, 1991, the Court held a hearing to determine Cornell’s fair share of the response costs incurred by the government at PAS. The Court adopted the six-factor fair share allocation test set forth in United States v. R.W. Meyer, Inc., 932 F.2d 568 (6th Cir.1991), and found Cornell liable for six percent of the response costs recovered from Alcan and, thus, the Court awarded Alcan $310,540.92. On appeal, the Second Circuit affirmed the grant of summary judgment with respect to the imposition of liability against Alcan for response costs at PAS and the finding that Alcan was entitled to contribution from Cornell. See United States v. Alcan, 990 F.2d 711 (2d Cir.1993). However, the Second Circuit reversed the Court’s finding that the government was entitled to summary judgment on the issue of damages. See id. In “essentially” adopting the reasoning of the Third Circuit in United States v. Alcan Aluminum Corp., 964 F.2d 252, 267-271 (3d Cir.1992) (“Alcan-Butlei•”), the court found that the common law scheme of joint and several liability applies to CERCLA. Under this rubric, where two or more tortfeasors act independently and cause a distinct or single harm, for which there is a reasonable basis for division according to the contribution of each, then each is liable for damages only for its own portion of harm. In other words, the damages are apportioned. But where each tortfeasor causes a single indivisible harm, then damages are not apportioned and each is liable in damages for the entire harm. 990 F.2d at 722 (citing Restatement (Second) of Torts § 433(A)(1965)). ■ The Second Circuit “candidly admitted” that its holding brought causation into CERCLA, a strict liability statute, through the back door, but limited the “special exception to the usual absence of causation” to situations where a defendant’s “pollutants did not contribute more than background contamination and also cannot concentrate.” Id. Thus, the court held: Alcan may escape liability if it either succeeds in proving that its oil emulsion, when mixed with other hazardous wastes, did not contribute to the release and the clean-up costs that followed, or contributed at most to only a divisible portion of harm. Id. (citing Alcan-Butler, 964 F.2d at 270). The court further instructed that Alcan “may present evidence relevant to establishing divisibility of harm, such as, proof disclosing the relative toxicity, migratory potential, degree of migration, and synergistic capacities of the hazardous substances at the site” to show that the harm at PAS was divisible. Id. (citing Alcan-Butler, 964 F.2d at 270 n. 29, 271; United States v. Monsanto, Co., 858 F.2d 160, 172 n. 26 (4th Cir.1988), cert. denied, 490 U.S. 1106, 109 S.Ct. 3156, 104 L.Ed.2d 1019 (1989)). Finally, the court held that if Alcan can show divisibility of harm they must also provide a “reasonable basis for apportionment of liability.” Id. At that time, the case United States v. Alcan Aluminum, 91-CV-1132 (“Alcan-Fulton”), involving clean-up costs at a su-perfund site in Fulton, New York was also pending before the Court. On December 1, 1993, the Court consolidated Alcan-Fulton with the PAS case. In May of 1996, the government moved for summary judgment against Alcan on two issues: (1) 'liability in the Alcan-Ful-ton case and (2) divisibility of harm and apportionment of costs in Alcan-PAS. Al-can opposed this motion, cross-moved for summary judgment on the same issues, and moved to dismiss for lack of subject matter jurisdiction, claiming that CERC-LA (1) cannot be applied retroactively and (2) is violative of the Commerce Clause. On October 28, 1996, the Court granted the government’s motion for summary judgment with respect to liability at the Alcan-Fulton site , denied Alcan’s motion to dismiss for lack of subject matter jurisdiction, and reserved decision pending re-briefing on the issues of divisibility and apportionment of harm. See United States v. Alcan Aluminum Corp., 1996 WL 637659 (N.D.N.Y. Oct.28, 1996). After considering the supplemental briefing submitted by both parties, the Court denied the government’s motion for summary judgment; finding that genuine issues of fact existed with respect to the constituents of Alcan’s emulsion; whether the emulsion contributed to the response costs and clean-up; and whether the metals in Alcan’s emulsion could have concentrated and, thus, posed an environmental threat. See United States v. Alcan, 1997 WL 727506, at *4 (N.D.N.Y. Aug.20, 1997). On October 29, 1998, Alcan moved to dismiss the Complaint pursuant to Fed. R. Crv. P. 12(b)(6), asserting that the retroactive application of CERCLA is unconstitutional in light of the Supreme Court’s decision in Eastern Enter. v. Kenneth S. Apfel, 524 U.S. 498, 118 S.Ct. 2131, 141 L.Ed.2d 451 (1998). The Court denied this motion in a Memorandum-Decision & Order dated May 11, 1999. See United States v. Alcan Aluminum Corp., 49 F.Supp.2d 96 (1999). Cornell was released from this ease by a stipulation signed September 29,1999. The Court held a bench trial in this action during October 4 — 8, 1999 to determine whether Alcan could avoid or limit liability. At trial, the government argued that Alcan could not avoid liability because: (1) Alcan could not prove that its emulsion, when mixed with hazardous substances, did not contribute to the release or clean-up costs or contributed to a divisible portion of the harm; (2) Alcan’s emulsion contained polychlorinated biphenyls (“PCBs”), Volatile Organic Compounds (“VOCs”), and above background levels of metals, which contributed to the release and clean-up costs; and (3) the metals in Alcan’s emulsion were capable of concentrating. Alcan, on the other hand, argued that: (1) its emulsion did not contain PCBs, VOCs, or above background levels of metals; (2) the metals did not concentrate; (3) that the legal standard adopted by this Court is incorrect; and (4) it could establish divisibility of harm by looking at the relative toxicity of hazardous substances at the PAS and Fulton Sites and demonstrating that the emulsion contained only background metals that did not concentrate. Alcan did not address the effect of its emulsion as a whole on the response costs at PAS and Fulton; instead, it focused on the individual constituents of its emulsion. II. Findings of Fact A. Alcan’s Plant and the Hot Rolling Process Alcan operates a hot mill, including a hot rolling line which processes aluminum ingots into sheets of aluminum. The process uses an oil and water emulsion to lubricate and cool the ingot and the steel rolls used to reduce the ingot’s width. The emulsion picks up fragments of the ingots during the rolling process. Although the emulsion is filtered before disposal, the filtration process does not remove all of the metal fragments. As a result, Alcan’s emulsion contains aluminum as well as the impurities in the ingot, cadmium, chromium, copper, lead, nickel, and zinc. Before the ingot is placed on the hot rolling line, it is heated in “pusher furnaces” (also called homogenizing furnaces or soaking pits). In the 1970s, Alcan cleaned the ingot with 1,1,1, Trichloroeth-ane (“TCA”) before placing it in these furnaces. After it is heated, the ingot is placed on the hot rolling line where it passes through an edging mill, two rolling mills, and various shears. The first rolling mill, the 120 inch reversing mill (the “120 inch mill”), reduces the thickness of the ingot by approximately one half inch each time the ingot passes through. Alcan uses a two to three percent oil and water emulsion as a lubricant and coolant in this mill (the “120 inch mill emulsion”). After the ends are sheared, the ingot travels to the 100 inch mill (also called the finishing mill or tandem mill) (the “100 inch mill”), which further reduces its thickness. An emulsion composed of about six percent oil in water is used to cool and lubricate the rolls in the 100 inch mill (the “100 inch mill emulsion”). The used emulsion is collected in sumps underneath each of the mills. The scum-like top layer of the emulsion is removed with a skimming system before the bulk of the emulsion is pumped into the system’s “dirty compartment” for storage. The emulsion is then sent through a belt type skimmer, and next, to a recirculation system. See Government Ex. 36, at p. 1. Distilled water is added to the emulsion to keep the concentration of oil and water constant. The emulsion is recirculated for up to one year before it is deposited in a waste emulsion collection system in the hot mill area and pumped to a central storage facility for processing and disposal. The sumps underneath the rolling mill are not continuous. There are walls or other barriers separating the 100 inch mill emulsion from the 120 inch mill emulsion. When the emulsions are sent to waste, however, they are combined. See Lagoe Testimony, Tr. at 168. At the time of disposal, the waste emulsion was composed of approximately eighty percent emulsion from the 100 inch mill, ten percent emulsion from the 120 inch mill, and ten percent hydraulic, lubricating, and bearing oils. See id. at 149. The floor and wall areas underneath the roll table and shear pits are cleaned only once per year; thus, the areas are not subject to a constant flushing of coolant during production. See Government Ex. 24. Until the mid-1980s, in addition to the hot rolling line, the hot mill area contained a plate mill (also called the flat sheet line and cut-to-length line). The plate mill converted the rolls of aluminum created by the hot mill into flat sheets of aluminum. Alcan used rags soaked with TCA to clean the rollers on this line. See- Lagoe Testimony, Tr. at 52. Other than the discontinuation of use of certain materials like TCA and fluid containing PCBs, there were no major changes on the hot rolling line from 1970 to 1989. In 1978, however, Alcan installed an ultra filtration unit. The ultra filtration unit processes the emulsion so that less material is sent out for disposal and more is recycled in house. See Lagoe Testimony, Tr. at 18. In the ultra filtration process, the emulsion is circulated through a membrane system, yielding two by-products: “the ultra filtration concentrate which is subsequently used as a fuel and a permeate which then flows to an on-site lagoon for further treatment.” See Government Ex. 36, at p. 2. Sometime in 1978, the ultra filtration unit was connected to outfall 002, which discharged treated permeate into Lake Ontario. See Lagoe Testimony, Tr. at 187; Government Ex. 3. B. PAS From 1970 to 1977, Pollution Abatement Services operated PAS, a high temperature liquid chemical waste incineration facility on fifteen acres in the City of Oswe-go, New York. Lake Ontario lies almost directly North of PAS and the entire site area slopes toward Lake Ontario. The site lies between two creeks (Wine Creek and No Name Creek), which intersect. The creeks collect surface water and carry it to Lake Ontario. See Appendix 1. A million gallon lagoon was located in the southern portion of the site, at the highest level of elevation. Leakage from the lagoon’s North side and overflow flowed North toward Lake Ontario. The middle area of PAS contained thousands of drums of “various chemicals in various conditions.” MacDonald Testimony, Tr. at 586. The lower portion of PAS contained a number of storage tanks and an incinerator. The incinerator stopped working at some point during PAS’s operation and waste that was meant to be incinerated was combined and stored in on-site lagoons. See MacDonald Testimony, Tr. at 585. Several liquid waste spills and lagoon overflows occurred while PAS operated the site. The combination of high rain fall and overflowing and leaking lagoons led to three oil spills in 1976. In April, the United States Coast Guard (“USCG”) responded to PAS to investigate an oil spill. Commander MacDonald, the investigating officer, found that the site was significantly contaminated. “Some of [the] drums were holding liquids. Some of the drums had completely disintegrated, leaving a jelly type of consistency that was intact. The ground was covered. The ground was — there were pools of contamination on the grounds. There was oil on the grounds. This facility had become a dumping ground.” MacDonald Testimony, Tr. at 586; see also Freestone Testimony, Tr. at 659; Yezzi Testimony, Tr. at 673; Government Exs. 86-54, 86-65 to 86-69, 86-84, 86-85. In response to the April 1976 spill, the USCG constructed a secondary lagoon (the “Coast Guard Lagoon”) on the Northern portion of the site to contain spilled waste, instituted a recycling system that returned run off to the million gallon lagoon, and employed absorbent beams to collect oil from the surface of the adjacent creeks. See Government Exs. 85, 125. In June of 1976, a second oil/chemical spill resulted from a combination of heavy rains, surface run-off, leaching from the lagoons, and spillage from the million gallon lagoon. The EPA’s testing confirmed a USCG report that the oil/chemical mixture included PCBs. See Government Ex. 85, at p. 2. In December 1976, a third spill occurred after which the EPA took control of the site and built a tertiary lagoon (the “EPA lagoon”) in a further attempt to contain run-off. In March of 1977, the situation at PAS became critical, due in part to heavy rain and water flow from the spring thaw. All of the on-site lagoons were in danger of overflowing and/or bursting; the influx of ■water from the spring thaw was greater than PAS’s discharge and/or diversionary-capacity; and there was a high chance of continued precipitation. See Government Ex. 125. Testing conducted at the behest of the -EPA indicated that the million gallon lagoon leaked and that the Coast Guard and EPA lagoons were potentially unstable. The State of New York constructed a drainage ditch to' divert uncontaminated surface run-off to Wine Creek. However, the pollutant saturated ground contaminated the ditch and further contaminated Wine Creek. Reducing the level of the million gallon lagoon was a primary concern, see Government Ex. 015, because its leakage contributed to the hazardous run-off and oily spills. See Yezzi Testimony, Tr. at 672. (The million gallon lagoon overflowed and material continuously leaked through the lagoon’s base and walls.); see also Government Ex. 62-004; 62-010; 62-011; 62-013; 62-030. The overflow and leakage from the lagoon flowed down through the barrel storage area prior to reaching the Coast Guard Lagoon. See MacDonald Testimony, Tr. at 610; Yezzi Testimony, Tr. at 688 (“Any material from the lagoon would come down what we, in this case, named the valley of the drums. You have the drums that were leaking and deteriorating and it would commingle with the waste coming down [from the million gallon] lagoon. We tried to put in some drainage ditches around to kind of control the — the spread of contamination [ ] along the surface of the site....”) Portions of the ground at the site contained so much contamination that “walking on the site was [to some degree like walking] on a wet sponge, a very wet sponge. To the North side of the lagoon wall, the area there was very, very saturated with contaminated materials.” Yezzi Testimony, Tr. at 678. Recovery wells dug to collect material flowing through the underground soil filled up rapidly with contaminated liquids. See Yezzi Testimony, Tr. at 690. Once the contaminants commingled, they could not be separated such that you could identify or treat an individual generator’s waste. See id. at 691. PAS abandoned the site in 1977, leaving approximately 12,000 drums of liquid waste and 100,000 gallons of bulk chemical waste on the site. Metals, VOCs, and PCBs were found in the lagoons and in drums removed from the site. Hazardous substances were released from the site into the soils, surface waters, stream sediments, and groundwater at and outside of the site. The results of-testing at PAS indicated that contamination was significant and widespread but nonuniform across the site. ,The soils, surface water, and sediments at PAS were contaminated with VOCs and metals and the soil was contaminated with PCBs. See Government Ex. 73. The recorded values of PCBs and VOCs found in the soil “suggest that contamination was a result of a multitude of separate sources since no clear areal distribution pattern of contamination was discovered.” See Government Ex. 73, at p, 4-16; see also Government Exs. 78, 79. The high level of metals found in the soil were generally consistent with both off-site samples and background soil levels for the eastern region of the United States. See, e.g., Government Ex. 72, 78, and 79. Nickel was found in PAS’s groundwater at concentrations that exceeded applicable clean-up and background levels. See Stipulated Fact No. 93; Government Ex. 79, at p. 8 (nickel appears to be the only site-related metal in the groundwater); Government Ex. 72, at pp. 58-60 (nickel was noteworthy because of its widespread appearance at high concentrations). At least some of the 4,607,380 gallons of hazardous liquid waste Alcan sent to PAS for treatment and disposal was stored in the million gallon lagoon. Alcan’s waste constituted approximately twenty-five percent of the total volume of waste sent to PAS. See Lagoe Testimony, Tr. at 230. Numerous trial witnessés testified that the million gallon lagoon contained three distinct layers of waste: “a 10-14 in. deep 300,000 [gallon] viscous oily layer on top, a 3-5 ft. deep 500,000 [gallon] aqueous layer which contained an intense blood-red colored liquid with a high solids content, a Total Organic Carbon (“TOC”) content of approximately 1.5% and high metal ion concentrations and a bottom layer of semi-solid sludge.” Government Ex. 125. Testing indicated that PCBs contaminated the lagoon’s top layer. See Government Ex. 62-025; 62-033; 62-051; 62-052; and 62-136; 84; and 85, at p. 2. Hydrochloride gas, which may facilitate the breakdown of an emulsion, was also disposed of in the lagoon. After taking control of PAS, the EPA contracted with Sealand Environmental Services to remove oil containing liquids from the million gallon lagoon and a 20,-000 gallon pit. Sealand vacuumed off the oily PCB contaminated layer of the million gallon lagoon and sent the waste for incineration. After exploring numerous alternatives, the EPA brought in a Mobile Physical Chemical Treatment unit to treat the aqueous layer of the lagoon. See Government Ex. 62-052 (the only alternative left for treating liquid wastes in an environmentally acceptable manners was to pass the full contents through a column). The carbon filtration unit (the “Blue Magoo”) used at PAS is a mobile physical chemical treatment system designed and built by Envirex in combination with the EPA. After conducting a “treatability study,” where Envirex determined “what kinds of inoculant [the government] should use to settle out suspended solids and dissolved metals,” and a “pilot plan study to determine what [the government’s] carbon utilization rate [and loading] should be,” the government processed waste water from the containment lagoons at PAS multiple times. See Freestone Testimony, Tr. at 652-53; see also Government Ex. 125. The government also used the Blue Magoo to treat the aqueous layer of the million gallon lagoon, which contained Alcan’s emulsion. Id. The liquid was treated with a process “consisting of PH adjustment, flocculation, mixed-media filtration and carbon adsorption,” Government Ex. 125, such that organics and suspended solids were removed. The organic materials included in the treated layer included oils. Id. Although the volume of organic materials in the treated layer led to a “very high carbon utilization rate,” the treatment cleaned the water substantially. Id. at 656. This treatment was part of a court ordered remedy at PAS. See Government Ex. 63. After the aqueous layer was treated, the sludge-like layer was removed and sent to a landfill site. Approximately fifty compressed gas cylinders were discovered in the lagoon when the sludge was partially removed. See, e.g., Government Ex. 85, at p. 7. Before the cylinders were disposed of, they were examined, cleaned, and when necessary, emptied. Testing of the sludge layer revealed the presence of .50 ppm cadmium; 31.00 ppm chromium; 136.00 ppm lead; and numerous VOCs. See Government Ex. 91, at Table I; Government Ex. 98, at Table 3-3. The clean-up of PAS continued from 1978 through 1981. Underground storage tanks and drums filled with hazardous materials, including PCBs were discovered and removed. In 1981, the surficial clean-up of PAS began. The clean-up included the demolition and disposal of on-site facilities, removal of approximately 8,000 remaining drums, and the drainage and disposal of approximately 80,000 gallons of liquid chemical waste from on-site bulk storage tanks. See Government Ex. 73, at 4-2. In 1982, oily waste continued to permeate PAS’s soils and surface water. In November of 1982, the New York State Department of Environmental Conservation (“DEC”) hired URS Company to perform a complete site investigation, characterize the remaining contamination, recommend remedial actions, and design the remedial program for PAS. Pollution reports filed by the EPA indicated that despite significant clean-up efforts, PAS still posed a serious environmental threat and in September 1983, PAS was placed on the National Priorities List. See 42 U.S.C. § 9605(a)(8)(b); 40 C.F.R. Part 300, Appendix B. In 1984, the EPA issued a Record of Decision (“ROD”) selecting the remedial action to be implemented at PAS. The ROD provided for: 1) containment of hazardous substances at PAS through construction of a soil bentonite slurry wall; 2) construction of an impermeable cap; 3) installation of a leachate collection system within the slurry wall; and d) installation of a groundwater recovery system and on-site treatment of the recovered leachate and groundwater. See Government Ex. 84. By 1986, the DEC completed all of its remedial activities other than the on-site treatment system. The EPA and DEC continued to monitor PAS and in 1990, the EPA issued an administrative order on consent to PAS’s potentially responsible parties (“PRPs”), other than Alcan, to perform a supplemental remedial investigation. In September 1993, (after considerable work had been done) the EPA issued a supplemental ROD for PAS to address outside contamination. A group of PRPs conducted a Supplemental Pre-Remedial Design Study at PAS. On August 10, 1998, the Court approved a Consent Decree between the EPA and 80 PRPs, which required the non-de minimis PRPs to implement the remedy in the Supplemental ROD and reimburse the EPA for $963,300 in past costs and the first $500,000 of future response costs. The de minimis PRPs paid the EPA their volumetric share of past costs, totaling $86,269, and agreed to pay their volumetric share of future work and response costs plus a premium. In September of 1997, the EPA issued a second supplemental ROD for PAS to address PCB contamination in adjacent creeks and wetlands. The ROD recommended no further action to remediate the contamination other than long term monitoring of PCB levels. C. Fulton Terminals From 1972-1977, Fulton Terminals, Inc., a subsidiary of PAS, operated a staging and storage area for materials awaiting incineration at PAS (“Fulton”). Fulton consists of approximately 1.6 acres in a light industrial/commercial area of the City of Fulton in Oswego County, New York. The Fulton Site housed five above-ground storage tanks, one partially buried storage tank, and two below-ground storage tanks. See Appendix 2. Essentially, Fulton housed waste that PAS could not accept because it was over capacity. The contamination at Fulton was due primarily to leaks in the storage tanks. See Rotóla Testimony. The Court previously determined that Alcan’s waste emulsion was stored at Fulton and that the emulsion was hazardous within the meaning of CERCLA. See Al-can, 1996 WL 637559, at *3, *15. The parties have stipulated that Alcan’s waste was stored in Tank One, which was an above-ground tank located in the site’s South West corner. See Appendix 2. Although sample analyses indicated significant PCB contamination at Fulton, tests of the waste remaining in Tank One found less than .01 |xg/g PCB. See Rotóla Testimony; Government Ex. 98, Table 3-1; see also Government Exs. 88-90. Tests of the soil and groundwater near Tank One revealed VOC and metal contamination. VOC contamination included vinyl chloride, TCA, and 1,1 dichloroethene, while heavy metal contamination included cadmium, lead, nickel, and chromium. See Rotóla Testimony. The metal contamination may have been naturally occurring as the sample results generally fell within typical values for soil reported in eastern regions of the country. See id.; see also Government Exs. 100, 101 (finding that in general, on-site metal findings are “well within typical values for the eastern region.”), and 102. The EPA listed both nickel and vinyl chloride as “contaminants of concern” at Fulton. See Rotóla Testimony, D. Possible Constituents of Alcan’s Emulsion From 1970-1977, Alcan sent 4,607,380 gallons of used emulsion to PAS. Between August 1972 and January 1973, Alcan shipped approximately 70,000 gallons of emulsion to Fulton. No samples of Alcan’s emulsion were taken prior to its disposal. Accordingly, the parties and the Court looked to samples of Alcan’s emulsion taken between 1981 and 1989; Alcan’s use and purchase of materials in its Oswego facility; the contamination of Alcan’s facility; and the results of government samples taken from PAS, Fulton, Butler Tunnel, and Sealand to determine the likely constituents of Alcan’s emulsion. Samples from Butler Tunnel and Sealand are relevant because Alcan sent waste emulsion to these sites in the late 1970s and early 1980s. Because no major changes occurred in Alcan’s hot rolling process between 1970 and 1989, the waste sent to Butler Tunnel and Sealand is representative of that sent to PAS and Fulton. 1. Metals This Court has previously determined, and the parties now stipulate, that, in addition to oil and water, Alcan’s emulsion contained aluminum, cadmium, chromium, copper, lead, and zinc. The metal in the emulsion resulted from fragments or “fines” of the ingot breaking away into the emulsion during the rolling process. Impurities in the aluminum ingot included nickel. There is no scientific reason to expect that nickel would behave differently from the other metals in the ingot. In other words, if the other metals made their way into the emulsion during the rolling process, there is no reason to expect that nickel would not have done the same thing. See Meyer Testimony, Tr. at 758; Driscoll. Alcan had the opportunity to test its emulsion for nickel, yet offered no evidence other than the absence of testing by the government to contradict testimony suggesting that the emulsion contained nickel. The Court is not aware of any sampling results that indicate nickel was not in the emulsion. On the basis of Dr. Meyer’s testimony and the opinion of the Court appointed expert, the Court finds that it is more likely than not that Alcan’s waste emulsion contained nickel. See id. 2. PCBs The government argued that the Court should find that PCBs contaminated Alcan’s emulsion in the 1970s, when Fulton and PAS were in operation, based on the totality of the circumstances. This argument works backwards from Alcan’s discovery of PCB contamination in its Oswego Plant in 1981 and is supported by the temporal circumstances surrounding the use and availability of PCBs and the extent of the contamination at Alcan’s plant. Alcan, on the other hand, argued that the emulsion sent to PAS and Fulton did not contain PCBs. Alcan’s argument is supported by the testing for PCBs conducted at PAS, Fulton, and Butler Tunnel and the absence of sampling data prior to 1981. The evidence shows that Alcan purchased hydraulic fluids and lubricants containing PCBs in the late 1960s and 1970 from Monsanto and E.F. Houghton & Co. Alcan used these fluids in the casting pits in the remelt area and in transformers in various area of the Oswego plant, including the recycling center, remelt area, the homogenizing furnaces, and the cold mill. Monsanto, the sole domestic manufacturer of PCBs, ceased production of PCB containing fluids in 1972 and, thus, PCB fluids were not commercially available in the United States after this date. See Kraft Testimony, Tr. at 317. In October of 1976, Congress passed the Toxic Substances' Control Act (“TSCA”), 15 U.S.C. §§ 2601 et seq. and the EPA promulgated implementing regulations, which applied to all types of hydraulic equipment. The regulations allowed continued use and topping off of systems containing less than 50 ppm PCBs through 1984; required the draining of systems containing more than 50 ppm PCBs; and prohibited the introduction of any PCB fluids in new hydraulic systems after 1979. See Kraft Testimony, Tr. at 315-16. Because PCB containing hydraulic fluids were not commercially available domestically after 1972 -and the TSCA and its implementing regulations strictly limited the percentage of PCBs in hydraulic equipment, the evidence suggests that PCB concentration in hydraulic equipment was lower in the 1980s than in the 1970s. See Kraft Testimony. In fact, in the. early 1980s, the EPA conducted testing of industrial hydraulic equipment to ensure compliance with the TSCA and the highest concentrations discovered measured around 3,000 ppm. See id. The evidence further suggests that Al-can complied with the TSCA and that an internal policy prohibiting the purchase of PCB materials or things known to contain PCBs curtailed Alcan’s use of PCBs. See Lagoe Testimony, Tr. at 183. Thus, the Court finds that the concentration of PCBs in Alcan’s hydraulic equipment mimicked that of industry in general insofar as the percentage of PCBs in the equipment was lower in the 1980s than -the 1970s. See Kraft Testimony, Tr. at 324. The results of internal testing of Alcan’s hydraulic equipment, which found minimal levels of contamination in two units in 1983 (2-3 ppm) and no contamination in 1987, support this conclusion. See Government Ex. 21. In June of 1980, the DEC detected PCBs in Alcan’s processed water discharge outfall 002. See Government Exs. 3, 6, Lagoe Testimony, Tr. at 188. Alcan subsequently investigated the source of the PCB contamination under DEC’s supervision and discovered contamination in the ponds and wetlands the water passed through en route to outfall 002. Alcan traced. this contamination to the use of PCB hydraulic oils in the remelt operation in the late -1960s and 1970. See Lagoe Testimony, Tr. at 244. Despite the unavailability of PCB fluids and strict guidelines on PCB use, in-house testing of Alcan’s waste emulsion conducted in 1981 showed PCB contamination. In 1983, testing revealed PCB contamination (1 ppm to 45 ppm range) throughout Alcan’s entire waste oil system. See Government Ex. 20. The -waste oil contamination resulted from contamination in the concrete below the hot roll mill — PCBs residing in the concrete leached into the emulsion. See, e.g., Lagoe Testimony, Tr. at 390-91. Between 1981 and 1987, Alcan conducted extensive in-house testing to determine the source and extent of PCB contamination and in 1988, Alcan solicited bids for a PCB remediation project. Testing of the contaminated area revealed extensive contamination in the concrete walls and floor of the hot line. O’Brien and Gere (“OBG”) conducted tests of the concrete up to a depth of 3 inches. The parties stipulate that the tests found consistent PCB contamination at depths up to three inches in Sections C and D of the roll table. See Government Ex. 23. Moreover, the concrete samples from Section D contained between 1,200 and 64,000 ppm PCBs and the average concentration of PCB contamination in the roll trench floor was 14,944 ppm PCBs. See id. Samples of the walls of the roll trench table found an average of 10,142 ppm PCBs. See id. Further PCB contamination was found on the roof of the hot line and on a crane located 30 to 45 feet above the hot line. The parties stipulated that PCB contamination extended to the height of the ceiling and crane because “fugitive dusts, mists, and liquids containing PCBs migrated from the Hot Line to these areas above the Hot Line.” Stipulated Fact No. 37. This data reveals extensive and widespread contamination in the hot roll mill that can only be explained by extensive use and/or spillage of PCB containing materials. According to an expert witness, approximately 100 to 200 gallons of PCB contaminated fluid would need to have been spilled to account for the contamination in the hot line’s floor alone. See Connolly Testimony, Tr. at 474; Lagoe Testimony, Tr. at 185. In this litigation, Alcan maintained that the in-house investigation did not conclusively determine the source of the concrete’s contamination. Alcan theorized that the contamination may have resulted from a spill from hydraulic equipment in 1983 near Section E of the hot mill as depicted in Appendix 3. However, Alcan did not report any spills of PCBs to New York State or the EPA in the 1980s or warn workers that there may have been a PCB spill during this time period. See Lagoe Testimony, Tr. at 184. Moreover, tests indicate that Alcan’s emulsion was contaminated in 1981, two years before the possible spill, see Government Exs. 8, 13, and that the relative concentration of PCB contamination in Section E was lower than Section A. See Connolly Testimony, Tr. at 477. Furthermore, in 1988, Alcan did not explain this theory of contamination to potential bidders on. the remediation project. Instead, Alcan told the bidders that “the probable source of contamination was its historical use of a PCB fluid on the roll table system.” See Government Ex. 23; Government Ex. 25. Employees or representatives of Alcan repeated this theory to bidders. See Government Ex. 25. Moreover, Lagoe testified that, during the relevant time period, the roll table was cleaned with emory paper and lubricated with oil once a week. See Lagoe Testimony, Tr. at 153. Significantly, Lagoe, the only Alcan witness to testify at trial, did not have personal knowledge as to what kind of oil was used to lubricate the roll table; Lagoe was not in charge of environmental compliance in the 1970s, did not supervise the cleaning of the roll table, and did not have an office in the direct vicinity of the table. See Lagoe Testimony. In combination with the extensive contamination of the concrete in the hot roll mill, the facts that: (1) Monsanto stopped producing PCB hydraulic oils in 1972; (2) the EPA implemented the TSCA regulations severely limiting and eventually prohibiting the use of PCBs in 1976; (3) testing identified significant levels of PCBs in Alcan’s waste oil and ultra filtration concentrate in 1981; (4) only two pieces of Alcan’s hydraulic equipment had minimal PCB contamination in 1983; (5) Alcan did not report any PCB spills to government agencies or its employees in the 1980s; (6) the relative concentration of PCB contamination in Section E was lower than Section A; and (7) Alcan explained the- probable source of contamination to bidders on the remediation project as historical use of PCB oils on the rolling table, the Court finds it more likely than not that Alcan’s theory that the concrete became contaminated because of a possible contractor spill by the 100 inch roll mill in 1983 is not credible and that the historical use of PCBs on the hot roll line caused the PCB contamination. Although there is evidence of significant contamination at Alcan’s facility, testing at Fulton, PAS, and Butler Tunnel did not identify significant contamination of PCBs in the areas where Alcan’s waste was deposited. This testing, however, does not necessitate a conclusion that .Alcan’s waste did not contain PCBs. A number of factors including, but not limited to, the variability of waste at the sites, conditions of the sites, material tested, and the number of samples taken impact the test results’ value. For example, only one of many samples taken at Butler Tunnel detected PCBs (.052 parts per billion), see Towle Testimony, Tr. at 556; Government Ex. 109. However, neither the actual waste disposed of at Butler Tunnel nor the oily discharge compelling clean-up action were tested for PCBs. There [were] probably a number of reasons why there [were] so few positive detections for PCB. Among them and most important would be that most of the samples are being collected at a time when there are not currently wastes being discharged from the Butler Tunnel. Second, there is a large amount of dilution that’s occurring, so whatever residual wastes that are left behind are being diluted with a lot of uncontaminated mine water also coming out of the Butler mine tunnel. And additionally, the PCB detection was of a water sample rather than an oil or waste sample, and PCBs are not readily soluble in water. Towle Testimony, Tr. at 556-57. In other words, the tests conducted at Butler Tunnel tested water discharge rather than the oily discharge or oily water. The fact that PCBs not readily soluble in water is not disputed. See, e.g., Lagoe Testimony, Tr. at 219 (slightly soluble); Towle Testimony, Tr. at 556-57. Accordingly, if PCBs were present in the oily discharge or waste, one would expect the majority to travel with the oily waste rather than remain in the water. Because of limits on testing and the testing conditions at PAS, it is equally problematic to conclude that the absence of detection of PCBs at certain locations, specifically the middle aqueous layer and bottom sludge-like layer of the million gallon lagoon, indicates the complete absence of PCBs. Early testing of the Lagoon did not include tests for PCBs because a standard procedure was not available to test for PCBs in a “mixed-organic/oil like matrix.” See Government Ex. 119. Later testing, however, indicated that PCBs contaminated the top oily layer of waste. See, e.g., Government Ex. 62. The fact that significant levels of PCBs were found in the top oily layer of the million gallon lagoon conclusively establishes that PCBs were present in the lagoon, despite the fact that they were not evenly distributed throughout the waste layers. Accordingly, it would be illogical to infer from tests of the middle and bottom layer of the lagoon, which did not find PCBs, that none of the waste deposited in the lagoon contained PCBs. Furthermore, based on the considerable variability of the waste, discussed below, and testimony regarding the condition of PAS in general, it is evident that no single generator’s waste remained in isolated or even recognizable form after disposal. Thus, the fact that Alcan initially disposed of liquid in an emulsion form does not mean that Alcan’s waste constituted only the middle layer of the lagoon. In fact, as discussed in Section 11(E), it is highly unlikely that Alcan’s waste retained its emulsion form after disposal and, thus, it is more likely than not that some of Alcan’s waste resided in the PCB contaminated oily layer of the lagoon. The considerable variability in the lagoon’s layers casts further doubt on the conclusiveness of PCB testing. Numerous witnesses testified that the contents of the million gallon lagoon had substantial variability both within and without the three identifiable layers of waste. See Freestone Testimony, Tr. at 651 (“[T]he sludge had variability, depending upon where you sampled.”); Yezzi Testimony, Tr. at 684-85; and MacDonald Testimony, Tr. at 584. The high variability of wastes suggests that different findings may have resulted at different sample locations. At Fulton, Alcan’s waste was disposed of in Tank One. At the time of the clean-up and testing, however, all that remained in Tank One was a sludge like material, which Potentially Responsible Parties (“PRPs”) eventually removed with shovels. While no PCBs were found in the sludge-like material remaining in the bottom of the tank, Alcan’s emulsion was not present either. Neither Alcan nor the government suggests that in the face of evidence that Alcan disposed, of emulsion and a stipulation that this emulsion was stored in Tank One, the Court should conclude that the absence of the emulsion at the time of testing suggests that the emulsion was never there. Accordingly, the Court will not conclude based on the absence of PCBs in Tank One’s sludge that the tank never contained PCBs. Seventeen of forty soil samples taken from the land spread area at Sealand tested positive for the presence of PCBs, as did two samples from the storage tank. See Government Ex. 115. As discussed above, Alcan’s waste emulsion constituted seventy-two percent of the material land spread at Sealand. Contrary to the sampling data at Butler Tunnel and Fulton, this data tends to suggest that Alcan’s emulsion contained PCBs. Based on the significant PCB contamination of Alcan’s hot rolling line that resulted from the historical use of PCB fluid on the rolling table and the facts that availability and use of PCB fluid was strictly curtailed, by the mid-1970s and PCBs in the concrete leached into Alcan’s waste emulsion, the Court finds that the government proved, by a preponderance of the evidence, that PCBs contaminated the emulsion Alcan sent to PAS and Fulton. See Meyer Testimony, Tr. at 771. 3. VOCs The government also contends that, at the time Alcan’s emulsion was disposed of at PAS and Fulton, it was contaminated with TCA. Again, there are no samples of Alcan’s emulsion from this period, so the parties and the Court looked to circumstantial evidence to determine whether Alcan’s emulsion contained TCA. The government argued that, based on Alcan’s admitted use and purchase of TCA, the TCA contamination in the hot mill area at Alcan’s Oswego plant, the EPA’s 1989 tests finding TCA in the emulsion, and the presence of TCA and its biological breakdown products at PAS, Fulton, Sealand and Butler Tunnel, the Court should find that TCA was present in the emulsion disposed of at PAS and Fulton. See, e.g., Meyer Testimony. Alcan responds that because TCA is highly volatile, it would have been impossible for TCA to get from the scalping area or plate mill, where they conceded use, into the emulsion. See Lagoe Testimony, Tr. at 63-64. Alcan further argued that the 1989 testing data is unreliable because the pages relating to TCA had “do not use” and the data had hand written changes and/or J qualifiers. See id. at 66-68. Alcan used TCA as a cleaner and de-greaser in the scalping area and plate mill at the Oswego Plant. See Lagoe Testimony, Tr. at 51 -52. Alcan also used TCA to clean the rolls in the gear box on the plate line. Alcan purchased TCA from Ashland Chemical Company. The use of TCA led to contamination of ground water at the plant. Alcan believes that the contamination resulted from a leaking sump adjacent to a scalper where TCA was used for cleaning purposes. See Lagoe Testimony, Tr. at 253. In 1978, Alcan had a TCA limitation on outfall 001, which the DEC implemented after finding TCA in the outfall. Outfall 001 included a waste stream from the hot line. See Government Ex. 2; Lagoe Testimony, Tr. at 197. Over time, TCA biodegrades. The biological breakdown products of TCA are 1,1 dichloroethane (“DCA”), 1,1 dichloroethy-lene (“DCE”), vinyl chloride, and chloroe-thane. See Government Ex. 128, at p. 363; Meyer Testimony, Tr. at 765-70. Tests conducted in 1989 identified VOCs, including TCA and two or three of TCA’s biological breakdown products DCA, chloroethane, and 1,1 Dichloroeth-ene in Alcan’s emulsion. See Government Ex. 36; Meyer Testimony, Tr. at 770. The fourth breakdown product, vinyl chloride, boils at a low temperature and, thus, would likely evaporate and escape into the atmosphere. See Meyer Testimony, Tr. at 770. Sampling at Butler Tunnel, Fulton, and PAS indicated the presence of TCA and its biological breakdown products. Sampling of the land spread area and storage tank at Sealand, however, did not find detectable concentrations of TCA. See La-goe Testimony, Tr. at 68; Nunes Testimony- Finally, the EPA’s sampling in 1989 indicated the presence of TCA in Alcan’s emulsion. Alcan argued that this data is unreliable because data pages contain hand written changes, J qualifiers, and/or state “do not use.” See Government Ex. 36. Kathleen Donnelly, the EPA employee who prepared the data in question, explained that the EPA’s standard operating procedures require all usable data in a report to be consolidated on one page. Thus, the usable data on sheets that read “do not read” was consolidated on a separate page. The data used in the report was “QA/QC’d,” which means that all of the EPA’s “standard operating procedures for insuring that the samples were analyzed and reviewed in accordance with the highest standards to make sure that the data [reported] is accurate.” Id. at 288. The J qualifier on data does not indicate that = it is suspect, but that a positively identified substance has an estimated concentration. J qualified results are reliable and usable. See Government Ex. 135; see also Cooper, 1996 WL 550128, at *15. Assuming all of the facts asserted by the government to be true, such that the 1989 tests reliably identified TCA and its biological breakdown products in Alcan’s emulsion; TCA and its breakdown products were detected at all four sites under review; Alcan used TCA near the hot rolling line during the relevant period; and Al-can’s groundwater and outflow 001 (which connects to the hot roll line) were contaminated with TCA during the relevant time periods, the Court cannot conclude that it is more likely than not that the emulsion sent to PAS and Fulton was contaminated with TCA. Unlike the evidence presented regarding PCBs, which proved that concrete at Alcan’s plant must have been contaminated by PCBs in the early 1970s and that this contamination leached into the waste emulsion and, thus, provided both a temporal and physical link between the PCB contamination and Alcan’s plant and the waste emulsion sent to PAS and Fulton, the evidence did not adequately link the TCA contamination, temporally or physically, with the emulsion sent to PAS and Fulton. E. Physical Characteristics and Interaction of Alcan’s Emulsion The emulsion Alcan sent to PAS and Fulton consisted of approximately 95% deionized mineral water and 5% Farbest mineral oil. The addition of an unknown emulsifying agent, which reduced the surface tension or interfacial tensions between the molecules of water and oil, created the emulsion. An “emulsion” is defined as “a suspension of small globules of one liquid in a second liquid with which the first will not mix, as milk fats in milk.” WEBSTER’S II NEW RIVERSIDE DICTIONARY (1994). In other words, in emulsion.form, oil and water exist in a single phase whereas in a simple mixture oil and water exist in separate phases. See Meyer Testimony, Tr. at 799. Thus, emulsions have different physical properties than either water or oil alone or a simple oil and water mixture. Id. To some degree, both water and oil dissolve hazardous substances. Some substances are more soluble in water than oil and some substances, like PCBs and VOCs, are more soluble in oil than water. An emulsion has the ability to dissolve larger concentrations of substances than either oil or water alone or a simple oil and water mixture and, thus, an emulsion increases the migratory potential of hazardous substances more than either water or oil. See Meyer Testimony, Tr. at 784; see also Government Ex. 129; Freestone Testimony, Tr. at 650 (“The normal solubility of most organic substances in water can be measured in the parts per billion and sometimes low parts per million range. For example, Fran Brezenski’s data for the Coast Guard lagoon shows various concentrations of contaminants in the parts per billion range. However, the intermediate layer in the main lagoon had a chemical bouillabaisse and such a high concentration of organic materials, in the Vk percent category. That’s 10,000 to 15,-000 parts per million. The only way you’re going to get that much organic material to stay in water is with some sort of mechanical or chemical emulsion process.”). Unlike oil, an emulsion does not float on top of water. Instead, when added to water, the emulsion will suspend and form a layer below the surface. However, an emulsion will not remain in its emulsion phase forever; instead, it eventually breaks down into its individual components. See Meyer Testimony, Tr. at 799; Driscoll. Exposure to the environment, heat, and acid increase the rate at which an emulsion breaks down. See Meyer Testimony, Tr. at 799; Driscoll. Therefore, it is unlikely that all of Alcan’s emulsion remained in the emulsion phase after disposal. Accordingly, the Court finds that it is more likely-than not that Alcan’s waste consisted of oil, water, and the water and oil emulsion. III. Conclusions of Law A. Legal Standard In the first stage of this action, the liability phase, the government had the burden of proving, by a preponderance of the evidence, that Alcan was liable under CERCLA. See infra note 2. The Court previously found that the government could introduce evidence at trial to prove that Alcan’s- emulsion contained PCBs, VOCs, and nickel. Pursuant to the standard outlined by the Second Circuit, if the government proved by a preponderance of the evidence that Alcan’s emulsion contained PCBs, VOCs, or above background levels of metals, Alcan could no longer escape liability altogether. See Alcan, 990 F.2d at 722 (“[C]ausation... is reintroduced only to permit a defendant to escape payment where its pollutants did not contribute more than background contamination and also cannot concentrate.”). As discussed above, after careful consideration of the evidence presented at trial the Court found that the government proved that Alcan’s emulsion contained PCBs, which are not naturally occurring substances that have “background levels.” Therefore, Alcan does not fit within the special category of defendants who can escape all liability under CERCLA Additionally, Acan is not able to escape all liability because the Court found that Acan’s emulsion contained nickel, which was detected at above background levels at both PAS and Fulton. This finding, combined with the expert testimony of Dr. Meyer and Dr. Driscoll, suggests that the metals in Acan’s emulsion can and, in this case, did concentrate. See Meyer Testimony; Driscoll letter. Acan did not negate this finding or present evidence sufficient to prove that the nickel in its emulsion can not concentrate. Athough Acan cannot escape liability altogether, it can limit its liability by showing that the harm at PAS and Fulton was divisible — •that is, it contributed, at most, to a divisible portion of the harm and that there is a reasonable basis to apportion liability for this harm. See Alcan, 990 F.2d at 722. To make this showing, Acan “may present evidence relevant to establishing divisibility of harm, such as, proof disclosing the relative toxicity, migratory potential, degree of migration, and synergistic capacities of the hazardous substances at the site.” Id. To satisfy its burden of proof, Acan must look at the potential effects of its emulsion as a whole rather than the individual components of the emulsion. Id.; see also Alcan, 1997 WL 727506, at *4 n. 5; United States v. Alcan Aluminum Corp., 892 F.Supp. 648 (1995), aff'd, United States v. Alcan Aluminum Corp., 96 F.3d 1434 (1996), cert. denied, 521 U.S. 1103, 117 S.Ct. 2479, 138 L.Ed.2d 988 (1997). Thus, as previously recognized by this Court, Acan “cannot succeed by arguing... that it avoids Lability simply because its metals [individually] are below background levels.” Alcan, 1997 WL 727506, at *4 n. 5 (internal citations omitted). Acan attacked this divisibility standard claiming that it is inconsistent with scientific principles, procedural and substantive due process, and Fed. R. Evid. 702. Acan argued that it can establish divisibility by comparing the relative toxicity of its emulsion’s individual components with the relative toxicity of other hazardous substances disposed of at PAS and Fulton. See Joint Pretrial Stipulation at 30. This argument, is based on the theory that one cannot scientifically evaluate the emulsion as a whole, but rather one must “look at the chemical impact of each of the individually chemically distinct parts.” Tr. of Sept. 10, 1999 proceedings at p. 12. On September 13, 1999, the Court denied Acan’s pre-trial motion to strike the legal- standard set forth by the Second Circuit and replace it with Acan’s proposed standard; the Court will not rehash the issue here. See id. It is sufficient to note, as the Third Circuit did when faced with the same argument, that: the government responds to ‘releases’ that threaten environmental safety. Thus, it is the release alone that must justify the response costs, not the particular waste generated by one given defendant. Here, there is no question but that a release occurred. Second, the fact that a single generator’s waste would not in itself justify a response is irrelevant in the multi-generator context, as this would permit a generator to escape liability where the amount of harm it engendered to the environment was minimal, though it was significant when added to other generator’s waste. Alcan-Butler, 964 F.2d at 264. Accordingly, in determining whether Acan proved divisibility, the Court will apply the standard set forth by the' Second and Third Circuits, which requires Acan to analyze its emulsion as a whole rather than focusing on the individual constituents of the emulsion. B. Divisibility The burden on Acan to show divisibility is substantial. See Alcan-Butler, 964 F.2d at 269. Courts are well aware that “[t]he practical effect of placing the burden on defendants [to prove divisibility] has been that responsible parties rarely escape joint and several liability. [Cjourts regularly [find] that where wastes of varying (and unknown) degrees of toxicity and migratory potential commingle, it simply is .impossible to determine the amount of environmental harm caused by each party.” See O’Neil v. Picillo, 883 F.2d 176, 178-79 (1st Cir.), cert. denied, 493 U.S. 1071, 110 S.Ct. 1115, 107 L.Ed.2d 1022 (1998) (citing United States v. Chem-Dyne, 572 F.Supp. at 811; Monsanto, 858 F.2d at 172-73). Courts recognize that the imposition of joint and several liability frequently results in one defendant paying more than its fair share of clean-up costs. Id. In the case at bar, the task of proving divisibility is complicated by the fact that neither the parties nor the Court know exactly what was disposed of at either PAS or Fulton or what was in Alcan’s emulsion. Although Alcan tangentially addressed “synergy” and “mixing” at trial, see, e.g., Lagoe Testimony, Tr. at 120-24, it did not provide evidence to negate the fact that its emulsion had a mobilizing effect on other hazardous wastes. As previously discussed, the Court found that Alcan’s emulsion increased the migratory potential of hazardous substances at PAS and Fulton. In so finding, the Court rejected Alcan’s analogy that the emulsion was synonymous to rain or milk added to cornflakes. Al-can’s analogy suggests that, like rain, Al-can’s emulsion contributed only volume to the waste at the sites. Alcan did not, however, prove that its emulsion was identical to rain in that it contributed only volume. Nor did Alcan prove that its emulsion was benign. In fact, the evidence suggested that Alcan’s emulsion was a more aggressive solvent than either water (rain) or oil (milk) because, in an emulsion phase, oil and water have a greater capacity to absorb contaminants than either water or oil alone. Alcan’s emulsion, therefore, absorbed the contaminants at the sites and facilitated their transport throughout. In so doing, Alcan’s emulsion contributed to the breadth of contamination at both PAS and Fulton. In other words, even if the hazardous substances in Alcan’s emulsion themselves did not contribute to the response costs at the sites (which the PCBs and nickel did), the emulsion as a whole contributed to both the release and response costs. Rain, unlike Alcan’s emulsion, would not have absorbed the quantity of organic substances, including VOCs and PCBs, which Alcan’s emulsion could have. The extreme conditions at PAS further complicate Alcan’s burden of proving divisibility. At least a portion of Alcan’s emulsion was dumped into the million gallon lagoon at PAS, which witnesses described as, among other things, a chemical bouillabaisse. See Freestone Testimony. Waste from the million gallon lagoon leaked through a compromised liner on the, bottom, pushed through the sides, and overflowed over the top of the lagoon. The leaking waste flowed throughout the site intermingling with, among other things, waste from leaking drums in the drum storage area. At the time the government commenced a clean-up, the ground was so saturated by liquid waste that, according to one witness, it was like walking on a wet sponge. As the Second Circuit recognized, however, the fact that Alcan’s waste commingled with other wastes at PAS and Fulton does not mean that the harm at the sites is indivisible. See Alcan, 990 F.2d at 722. The Second Circuit adopted the Restatement (Second) of Torts’ definition of divisibility, which states that harm is divisible where “joint tortfeasors act independently and cause a distinct harm, for which there is a reasonable basis of division according to the contribution of each.” Pros-ser and Keeton explain that where two or more causes combine to produce a single indivisible result, liability cannot be apportioned. See Prosser & Keeton on Torts, § 52, p. 347 (5th Ed.1984). Further, the conduct of multiple tortfeasors-need not.be simultaneous to cause an indivisible harm, “[o]ne defendant may create a situation upon which the other may act later to cause damage.” Id., at 348. For example, “[i]f two defendants, struggling for a single gun, su