YANKEE GAS SERVICES CO. v. UGI Utilities, Inc.

MEMORANDUM OF DECISION MARK R. KRAVITZ, District Judge. In this action, Yankee Gas Services Company (‘Yankee Gas”) and The Connecticut Light and Power Company (“CL&P”), both of which are subsidiaries of Northeast Utilities, sue UGI Utilities, Inc. (“UGI”) under the Comprehensive Environmental Response, Compensation, and Liability Act of 1980, 42 U.S.C. § 9601 et seq. (“CERCLA”), seeking to impose liability on UGI for pollution that occurred at thirteen facilities in Connecticut that produced manufactured gas from 1884 to 1941. These facilities are known as manufactured gas plants, or MGPs. At various points in time, CL&P was a subsidiary of UGI, or corporations in which UGI owned stock, until UGI was required to divest itself of its interest in CL&P in 1941 as a consequence of the passage of the Public Utility Holding Company Act of 1935. 1 The U.S. Supreme Court’s decision in United States v. Bestfoods, 524 U.S. 51, 118 S.Ct. 1876, 141 L.Ed.2d 43 (1998), provides the analytical framework for determining UGI’s liability for contamination at these sites. This is not the first time that courts have addressed UGI’s liability for pollution occurring at manufactured gas sites owned by its former subsidiaries, although this is the first case involving CL&P. See, e.g., Atlanta Gas Light Co. v. UGI Utilities, Inc., No. 3:03-cv-614-J, 2005 WL 5660476 (M.D.Fla. Mar.22, 2005), aff'd, 463 F.3d 1201 (11th Cir.2006) (“Atlanta Gas ”) (no liability for UGI under CERCLA); Consolidated Edison Co. of New York, Inc. v. UGI Utilities, Inc., 310 F.Supp.2d 592 (S.D.N.Y.2004), aff'd in relevant part, 153 Fed.Appx. 749 (2d Cir.2005) (“Consolidated Edison ”) (same). The Court and the parties agreed to try this case in two phases. The first phase was focused on certain statute of limitations issues as well as UGI’s derivative and direct liability within the meaning of Best- *232 foods. The second phase was to focus on the allocation of remediation costs between the various potentially responsible persons. The first phase of the case was tried to the Court over four days. Plaintiffs submitted 529 exhibits; UGI introduced 1,147 exhibits. Given the years at issue, the Court was impressed by the wealth of documentation that is still available, including most of the minutes of the relevant companies’ board meetings and committees and a substantial amount of correspondence and reports. 2 Because no individuals who actually worked at these sites during the relevant period were still available to be called at trial, the witnesses were retained experts and one fact witness regarding Plaintiffs’ clean-up efforts. The experts (none of whom was a historian) reviewed volumes of historical documents and gave their opinions based upon those documents or the impressions they derived from reading the documents. Before, during, and after trial, UGI sought to strike Plaintiffs’ experts on the basis of the Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), line of cases. The Court declined to strike the experts’ testimony, and so it remains in the record for what it’s worth. Truth be told, with a single exception, the Court did not find the experts’ testimony of much assistance in resolving the issues in this case. The Court says this for two principal reasons. For one, the Court was just as able as the experts to review the historical record and apply the Bestfoods standard to the facts as the Court found them. The Court did not need the experts to undertake that important task for the Court. For another, two of Plaintiffs’ witnesses — Dr. Neil Shifrin and Mr. Thomas Blake — appeared to analyze the historical record using standards that are contrary to those set forth in Bestfoods. For example, Dr. Shifrin declined to accept the possibility that a person having a position with a subsidiary and also with the parent corporation could act for the subsidiary while engaged in the subsidiary’s business. Dr. Shifrin’s presumption is contrary to the statement in Bestfoods that “courts generally presume that the directors are wearing their ‘subsidiary hats’ and not their ‘parent hats’ when acting for the subsidiary....” Bestfoods, 524 U.S. at 61, 118 S.Ct. 1876 (quotation marks omitted). Mr. Blake testified that when a company owned more than fifty percent of another company, that created a presumption (albeit a rebuttable one) that the parent controlled the subsidiary’s operations within the meaning of CERCLA. Mr. Blake’s presumption is contrary to the statement in Bestfoods that “the exercise of the control which stock ownership gives to the stockholders ... will not create liability beyond the assets of the subsidiary,” and that “nothing in CERCLA purports to reject this bedrock principle.” Id. at 61-62, 118 S.Ct. 1876 (quotation marks omitted). In effect, Mr. Blake improperly shifted the burden of proof to UGI. Where the Court did find the experts’ testimony of assistance (and did use it) was in connection with an understanding of the operation of the manufactured gas plants and the pollution they may have generated, as well as in interpreting technical aspects of the documentation. By closing arguments, certain issues had dropped from the case. Before trial, Plaintiffs withdrew claims relating to three sites — Waterbury Benedict Street, Meriden South Colony Street, and Winsted Prospect Street — without prejudice to re *233 newing those claims at an appropriate point in time. Also, at closing arguments, Plaintiffs withdrew their claim that the Court should pierce the corporate veil between UGI and CL&P — that is, in the parlance of Bestfoods, Plaintiffs withdrew their claim of derivative CERCLA liability. See Bestfoods, 524 U.S. at 64, 118 S.Ct. 1876. That concession focused the Plaintiffs’ claims solely on UGI’s direct liability as an “operator” of the remaining ten facilities. In particular, Plaintiffs sought to show that UGI operated the facilities “in the stead of [CL&P] or alongside [CL&P] in some sort of a joint venture.” Id. at 71, 118 S.Ct. 1876. UGI, for its part, agreed that it had operated the Waterbury North facility for the years in question but reserved its right to argue that Plaintiffs had not shown that any contamination occurred at that site during the relevant period of UGI’s operation. Accordingly, the Waterbury North facility was removed from the first phase of the trial and will be addressed, if need be, during the next phase. Thus, the facilities at issue in this first phase included the following nine MGPs: Norwalk, Bristol, Meriden Cooper Street, Middletown, Putnam, Rockville, Waterbury South, Winsted Gay Street, and Willimantic. Having carefully considered the documentary evidence and cognizant of the Bestfoods standard, the Court concludes that Plaintiffs have not sustained their burden of proving by a preponderance of the evidence that UGI operated these nine MGPs in the stead of CL&P or in some sort of joint venture with CL&P. Plaintiffs succeeded in showing that UGI was a vigilant parent that conducted detailed — yet not eccentric — oversight of the operations of its subsidiaries in Connecticut. But the facilities in question were operated by CL&P and its employees. To be sure, UGI provided assistance to CL&P from time to time when CL&P requested it; UGI also carefully oversaw the operations of CL&P, consistent with UGI’s status as a corporate parent. But that assistance and oversight is a far cry from managing, directing, or operating the facilities in the stead of CL&P or in some sort of joint venture with it. In short, from the documentary record provided, the Court believes that CL&P was accurate and truthful when it told the Securities and Exchange Commission .in 1940 that UGI “has never undertaken to control the affairs of this Company in any way.” Ex. 1586 at 850. In addition, and in the alternative, for at least two of the facilities, the Court finds that Plaintiffs’ cost recovery actions are time-barred. Before turning to the Court’s findings of fact and conclusions of law, the Court pauses to express its gratitude to counsel for both sides in this dispute. They were consummate professionals who dealt with the' Court civilly and with commendable candor at all times. Given the volume of evidence provided to the Court, it would have been impossible to review it meaningfully without the superb assistance from all counsel. I. What follows in this section are the Court’s findings of fact in accordance with Rule 52 of the Federal Rules of Civil Procedure. The Court will begin with a general overview of manufactured gas production and the companies in question. Given the complexity of the issues involved and in order to avoid unnecessary repetition, the Court will keep this first section relatively brief and will make further findings of fact in connection with the Court’s discussion of the various issues presented by the parties. Basic Operations of Manufactured Gas Plants. During the heyday of manufactured gas production' — between 1816 *234 and 1960 — approximately 1,000-1,500 MGPs were built and operated in the United States. See Ex. 1 at 9. Prior to the development of natural gas supplies and transmission systems during 1940s and 1950s, virtually all fuel and lighting gas used in the United States was manufactured at MGPs using one of three basic processes: (1) coal gas, which was produced from thermal destruction of coal in the absence of air; (2) oil gas, which was manufactured by enriching super-heated air with carbureted petroleum; and (3) carbureted water gas (“CWG”), which was made by blowing steam through red hot coke and/or coal to create water gas, followed by spraying (carbureting) petroleum into the hot water gas to enrich it. Id. CWG was the dominant process used in Connecticut (and nationwide) after the 1890s and is the process for which UGI had an exclusive patent. [[Image here]] Both tar and oil — referred to today as “non-aqueous phase liquids” or “NAPLs” — were generated as byproducts from all three manufactured gas processes. See id. at 10. Other byproducts included wastewater, purifier wastes, ash, clinker, and coke. See id. at 12. The amount of byproducts produced was substantial. Tar, in particular, was produced by the millions of gallons over the lifetime of an MGP. See id. at 10-11. Many of these byproducts were valuable in their own right and were collected and sold by the MGPs. For instance, tar was known as “black gold” and was sold for use as fuel, as well as for roofing, wood treating, road surfacing, dyes, medicines, disinfectants, explosives, and chemicals such as benzene or naphthalene. See id. at 17. Ash and clinker were also sold, mostly as fill and surface coverings for walks and driveways, Coke, the ideal “smokeless coal,” was sometimes sold as fuel. The recovery and sale of these byproducts was essential to the operation of MGPs and often proved more profitable than gas manufacturing itself. Thus, most MGPs had facilities for *235 separating tar and other byproducts and some had their own refining capabilities. See id. at 19. However, as both Plaintiffs’ and UGI’s experts testified at trial, it was difficult to operate MGPs without some leakage of tar and oil into the environment. Although MGPs attempted to minimize leakage (not because of the then-unknown environmental consequences but because of the monetary value of the byproducts), tar and oil could — and often did — enter the environment during many stages of the CWG process. In addition, catastrophic releases into the environment sometimes occurred, such as the release of oil and tar into the Norwalk River in 1926. See Ex. 147. Leakage also routinely occurred as a consequence of the demolition of former MGPs, although Dr. Shifrin acknowledged that there is no documentary evidence that UGI directed or managed the demolition of any Connecticut MGPs. See Tr. Transcript at 117. These various forms of contamination remain hazardous for decades or even centuries after entering the environment, and it is this contamination that Yankee Gas and CL&P, as current owners of the MGP sites at issue, are in the process of cleaning up. UGI’s History. UGI played a key role in the establishment of CWG as the prominent process for gas manufacturing in the eastern half of the United States. The CWG process was invented in 1872 by Thaddeus Lowe, a former consultant to President Lincoln and the chief of the Corps of Aeronautics. See Ex. 117 at 5721. Mr. Lowe’s new method produced gas with more illuminating power and at less cost than coal gas processes. See id. at 5722. In order to maximize the profit potential of the CWG process, Mr. Lowe teamed with other businessmen and investors to form the United Gas Improvement Company (UGIC) in 1882. Id. In agreements dating between 1882 and 1885, Mr. Lowe sold the exclusive right to sell, manufacture, and install the CWG process to UGIC and served as the chief engineer of the new company. See Ex. 107 at 5287. UGIC discovered that one of the most lucrative ways to profit from the CWG process was to buy stock in local public utility companies, advance them the money to implement the CWG process, and then make profit in the form of dividends. See Ex. 117 at 5726. However, like most corporate charters of the day, UGIC’s charter did not allow it to own stock in other companies. After several years of trying to circumvent this restriction through the use of trusts, UGIC formed another corporation called the Union Company, which purchased the corporate charter of the Union Contract Company. This charter had been granted to the Union Contract Company by a special act of the Pennsylvania legislature in 1870 and allowed the purchase and ownership of stock in other corporations. See Ex. 107 at 5289. In 1888, the Union Company changed its name to The United Gas Improvement Company (“TUGIC”), which later became UGI. 3 In 1889, UGI acquired UGIC’s property and assets, including the Lowe CWG patents. See Ex. 117 at 5726-27. The purchase of the Union Company’s charter allowed Pennsylvania-based UGI to acquire public utility companies in other states, including Connecticut, in order to profit from the installation of the CWG *236 process. UGI soon expanded beyond gas into the electricity industry, which was poised to replace the gas industry as the dominant energy supplier in the United States. See Ex. 109 at 5228. By the early 1900s, UGI had moved past simply selling and promoting the CWG process, investing not only in gas companies and electric utilities, but trolley, railway and lighting companies as well. See Ex. 117 at 5728; Ex. 1049. In addition to acquiring an interest in other companies, UGI also directly leased and operated some MGPs, particularly in its home state of Pennsylvania. By 1902, UGI held interests in 45 companies that were providing gas and electric services across the country. See id. at 5745. While UGI’s investment strategy in the early twentieth century consisted of piecemeal acquisitions of smaller companies, UGI began to consolidate its acquisitions in the 1910s and 1920s. UGI would acquire several gas and electric companies in one market, which it would then consolidate under a public holding company in which UGI owned a majority stake. See Ex. 109 at 5229. Thus, consolidated public utility companies were born in New Jersey, Indiana, Connecticut, and Pennsylvania. See Ex. 117 at 5759. In the 1920s, UGI continued to diversify, entering the construction field through its subsidiary, the UGI Contracting Company. See Ex. 107 at 5301. In 1928, UGI created United Engineers and Constructors (“UE&C”), under which it consolidated the UGI Contracting Company and several other international construction firms. In its first year of existence, UE&C did $68 million worth of business, including projects in Argentina, Brazil, Hawaii, and Europe. See Ex. 117 at 5764. The company built steam-power plants, electric generating stations, railroad passenger and freight terminals, hospitals, prisons, offices, and apartments. More than half of its work was done for companies not associated with UGI. See id. UE&C’s business peaked in 1929, shortly before the Great Depression, and UGI sold its stake in the company in 1938, refocusing its investments in the public utility sector, which was less devastated by the Great Depression than many other industries. See id. at 5766. By 1935, UGI had acquired an interest in public utility companies in Arizona, Connecticut, Illinois, Indiana, Michigan, New Hampshire, New Jersey, Ohio, and Tennessee. In some cases, UGI’s subsidiaries were wholly owned; in other eases UGI owned a majority or minority interest in the companies. See Def.’s Post-Trial Brief [doc. # 134], Ex. 4. UGI’s investment in its subsidiaries was in addition to its direct operation of many MGPs across the country. At its financial peak in late 1940, UGI had assets of $846 million, was operating in 11 states, and held investments in four sub-holding companies, 38 electric and gas utilities, and 48 non-utility companies, including water, transit, ice, and cold-storage companies. See Ex. 117 at 5716. UGIC and later UGI entered the Connecticut public utility market in 1894, when it leased the Waterbury North MGP from the Waterbury Gas and Light Company. See Ex. 327 at 39813. In 1900, UGI acquired the majority of the voting stock of the Connecticut Lighting and Power Company (a different company from CL&P), which changed its name in 1901 to the Connecticut Railway and Lighting Company (“CR & L”). See Ex. 226. CR & L owned the Norwalk Gas Light Company, which owned and operated the Norwalk MGP. CL&P later acquired the leases to CR & L’s properties. See Ex. 1006 at 3611. UGI’s involvement with CL&P’s predecessor, the Rocky River Power Company (“RRPC”), began in 1917 when the Con *237 necticut Electric Syndicate (“Syndicate”) acquired a majority interest in RRPC as well as several other public utility companies. These companies were then acquired by RRPC, which changed its name to CL&P. See id. UGI owned 80% of the Syndicate, having organized it in 1916 for the purpose of consolidating the public utility market in Connecticut. See id. In 1925, the Connecticut Electric Service Company (the “Service Company”) was organized and acquired all of the assets of the Syndicate in exchange for the issuance of common stock to the Syndicate. See id. Accordingly, CL&P became a subsidiary of the Service Company, which in turn became a subsidiary of the Syndicate. In February 1929, the Connecticut Electric Securities Company (the “Securities Company”) was organized and acquired all the assets of the Syndicate, which was then dissolved. Later that year, the Securities Company was merged into the Service Company, with CL&P as its subsidiary. By 1930, UGI owned 60% of the Service Company, which in turn wholly owned CL&P. See id. The Service Company was merged into CL&P in 1935, at which point stockholders of the Service Company, including UGI, became direct stockholders of CL&P. See id.; see also Ex. 327 at 39657-66. UGI divested its interest in CL&P in 1941. See Ex. 1006 at 3610. History of CL&P. CL&P was founded in 1905 as the Rocky River Power Company (“RRPC”). See Ex. 1012 at 10819. RRPC’s original charter only allowed the new company to utilize the power resources of the Rocky River, a small tributary of the Housatonic River. See id. However, RRPC’s founder, J. Henry Roraback, had plans to create a unified power system in western Connecticut. See Ex. 1009 at 1. He lobbied the Connecticut legislature for an expanded charter, which he obtained in 1909 and which brought him closer to accomplishing that goal. See id. The new charter allowed RRPC “to distribute power wholesale throughout the state, to construct dams on a stretch of the Housatonic in northwestern Connecticut, and to build mills and manufacturing plants in the same area.” Id. Yet, despite Mr. Roraback’s grandiose plans, he lacked the financial resources to carry out his unification plan until 1917, when RRPC was acquired by the Syndicate, a subsidiary of UGI. See Ex. 1012 at 10819; Ex. 1006 at 3611. Mr. Roraback is the central figure in CL&P’s- early history. A powerful person in Connecticut politics, Mr. Roraback was the chairman of the Republican Party in Connecticut from 1912 to 1937. See Ex. 1009 at 1-2. Although he never ran for political office, his political influence was legendary. See Ex. 1010 (“For 16 years Republican Boss John Henry Roraback ruled Connecticut.”). According to CL&P’s corporate history, Mr. Roraback’s “vision, judgment and ability were the key factors in the founding and successful expansion of the Company.” Ex. 1012 at 10821. For instance, in 1911, he drew a power map of the state that was so well conceived that when the power grid was built years later, it matched his original plans almost exactly. See Ex. 1011 at 1. Today he is hailed as the “builder and founder of the CL&P system.” Ex. 1012 at 10821. In 1917, with financial assistance from UGI, RRPC acquired the property and franchises of several other public utility companies in Connecticut and changed its name to CL&P. See id. UGI advanced CL&P $10 million in cash for these acquisitions and made it possible for the company to obtain credit for another $3 million. See Ex. 1012 at 10820. By 1921, CL&P had repaid its debts to UGI and thereafter financed its requirements through bond sales and loans from entities other than UGI. See id. Mr. Roraback became the *238 vice-president and director of the newly consolidated CL&P, later serving as president of the company from 1925 until his untimely death in 1937. See Ex. 1566. Other important figures in CL&P’s early corporate history include C.L. Campbell, who served as a CL&P director from 1923 to 1950 and as president of the company from 1937 to 1948; I.W. Day, who served as a director and vice-president from 1917 to 1939; and R.H. Knowlton, who served as a vice-president of the company from 1927 to 1948. See Ex. 1004 (summarizing biographies of CL&P directors and officers). CL&P had become the largest public utility in Connecticut by 1921 and continued to grow steadily in the 1920s and 1930s. See Ex. 1009 at 2. For example, while the company had only 43,436 customers in 1917, by 1937 CL&P had grown to having 160,790 customers and more than $15 million in revenue. See Ex. 1011 at 2. The company had also grown to 2,300 employees. See id. CL&P continued to acquire other public utility companies, and by 1941, it owned and operated the properties of 57 former public utility companies in Connecticut. See Ex. 1006 at 3611. It was during this time period that CL&P acquired a direct or indirect interest in (or in some cases became the sub-lessee of) the nine MGPs in question: Norwalk (1917), Meriden Cooper Street (1926), Bristol (1927), Middletown (1927), Putnam (1929), Waterbury South (1931), Winsted Gay Street (1931), Rockville (1935), and Willimantic (1935). 4 UGI Oversight. UGI oversaw its investments in Connecticut, including CL&P, in a number of ways, each of which the Court will say more about later in its section detailing its conclusions of law. First, UGI maintained a presence on the Board of Directors of the various holding companies that owned CL&P and on CL&P’s own Board of Directors. CL&P’s Board kept detailed minutes, which were admitted into the record. See Ex. 1125-1204. The Board also created a series of committees and sub-committees, such as CL&P’s Executive Committee, that were responsible for managing all aspects of its business. See, e.g., Ex. 1288, 1289, 1300, 1318, 1409. UGI never represented a majority of the directors on these Boards. See Ex. 1013,1014. UGI also kept track of its investments— in Connecticut and elsewhere — through a series of oversight committees — the Managing Committee and Works Committee (1889-1904), the Executive Committee (1904-1923), the Operating Committee (1923-1927), the Management Committee and the Committee on Operation and Maintenance (1927-1929), and the Second Executive Committee (1929-1943). See Ex. 1074, 1078, 1118, 1076, 1329. These committees provided recommendations to the subsidiaries’ boards, shared UGI expertise and best practices, and approved budgets and large capital expenditures. UGI also conducted (or least aspired to conduct) periodic financial audits of its subsidiaries, surveying accounting records, verifying cash, confirming collections, and verifying inventories. See Ex. 119 at 5604. In addition, UGI held annual Superintendent Conferences where representatives from UGI’s subsidiaries would gather to discuss the manufactured gas business and participate in an “exchange of ideas.” Ex. 2152. The agenda for these conferences included commercial topics such as the solicitation of new business and engineering topics such as research and development. See Ex. 1638. The annual conferences resulted in a series of “Engineering and Operating Notes,” in which best practices were outlined and local practices were com *239 pared. See Ex. 2059, 2060, 2155. As Dr. Shifrin admitted at trial, UGI began publishing its “Operating Notes” for industry-wide distribution as early as 1911. See Tr. Transcript at 67. Finally, UGI provided consulting services to its subsidiaries. Prior to 1927, UGI provided these services to CL&P and other subsidiaries on an informal basis, but in 1927, UGI and CL&P entered into a Management Services contract for the provision of consulting services. See Ex. 409 at 39889. This contract was similar to those that UGI entered into with other subsidiaries, except that CL&P’s fee was capped originally at $60,000 a year and later reduced to $20,000 a year. See Ex. 1517. Under the Management Services contract, UGI offered CL&P a wide variety of services, including accounting, legal, purchasing, and public relations services. See Ex. 1052 at 18842. According to the terms of the contract, UGI provided these services to CL&P when requested and always remained under the supervision of CL&P’s Board of Directors. See id. at 18864. There is evidence that CL&P did avail itself of UGI’s services, at least in certain cases. For instance, UGI negotiated bulk oil and gas purchasing agreements for its subsidiaries. See, e.g., Ex. 120 at 12919. UGI’s Divestiture from CL&P. By the late 1920s, the gas and electric industry had been largely consolidated into several sizable holding companies, of which UGI was among the biggest. This concentration of power, as well as some well-publicized abuses committed by public utility holding companies, caught the attention of the Federal Trade Commission (“FTC”) and Congress. This scrutiny ultimately resulted in a 1928 Senate resolution requiring the FTC to prepare detailed reports on “the actual investments, financial transactions, business operations, earnings and other economic results of electric power and gas utility companies, their holding companies, and their associated or affiliated companies.” Ex. 157 at 38050. The FTC conducted its investigations and issued several reports on UGI and its subsidiaries, including reports authored by Edwin T. Harris, see Ex. 157, and Lewis G. Pritchard, see Ex. 160. These reports led to Congressional hearings and finally, the passage of the Public Utility Holding Company Act of 1935 (“PUHCA”). See 15 U.S.C. § 79b(ll)(b)(l) (repealed Aug. 8, 2005). The PUHCA limited public utility holding companies’ operations to a single regional utility system which, in UGI’s case, consisted of the tri-state area of Delaware, Pennsylvania, and northern Maryland. See United Gas Improvement Co. v. SEC, 138 F.2d 1010, 1015 (3d Cir.1943). Under the supervision of the Securities and Exchange Commission (“SEC”), UGI was required to sell its subsidiaries in nine states, including Connecticut. UGI divested from CL&P in April 1941. See Ex. 1006 at 3610. By 1953, UGI was no longer a holding company and had instead become an operating company whose operations were limited to eastern Pennsylvania. Thus, with the exception of Norwalk, for which there is a dispute about an earlier time period of potential control, Plaintiffs’ claims are limited to facilities that were acquired between 1917, when UGI acquired an indirect majority interest in CL&P, and 1941, when UGI divested from CL&P. The Court now turns to these claims. II. Congress enacted CERCLA in 1980 to promote the “timely cleanup of hazardous waste sites and to ensure that the costs of such cleanup efforts were borne by those responsible for the contamination.” Burlington Northern & Santa Fe Rwy. Co. v. United States, U.S. -, 129 S.Ct. *240 1870, 1874, 173 L.Ed.2d 812 (2009) (citing Consolidated Edison, 423 F.3d at 94) (quotation marks omitted). To that end, CERCLA imposes strict liability for environmental contamination upon several broad categories of potentially responsible persons, one of which includes the “owner and operator” of the facility. See 42 U.S.C. § 9607(a)(1). As explained above, Plaintiffs seek to hold UGI liable for a portion of the clean-up costs as a former “operator” of the nine MGPs at issue. A. In United States v. Bestfoods, 524 U.S. 51, 118 S.Ct. 1876, 141 L.Ed.2d 43 (1998), the Supreme Court set out to answer the following question: “whether a parent corporation that actively participated in, and exercised control over, the operations of a subsidiary may, without more, be held liable as an operator of a polluting facility owned or operated by the subsidiary.” Id. at 55, 118 S.Ct. 1876. 5 The Supreme Court answered that question “no, unless the corporate veil may be pierced.” Id. However, the Court also held that a “corporate parent that actively participated in, and exercised control over, the operations of the facility itself may be held 'directly liable in its own right as an operator of the facility.” Id. The case before this Court turns on what it means to “actively participate[ ] in, and exercise[ ] control over, the operations of the facility itself.” Fortunately, the Supreme Court in Bestfoods gave lower courts considerable assistance in applying that principle, and since that guidance is so central to resolution of the issues in this case, the Court has taken particular pains to set it out at great length. It is well to recall that in Bestfoods, the Supreme Court began its analysis of parental liability with this statement: “It is a general principle of corporate law deeply ingrained in our economic and legal systems that a parent corporation (so-called because of control through ownership of another corporation’s stock) is not liable for the acts of its subsidiaries.” Id. at 61, 118 S.Ct. 1876 (quotation marks omitted). Thus it is hornbook law that the exercise of the control which stock ownership gives to the stockholders ... will not create liability beyond the assets of the subsidiary. That control includes the election of directors, the making of bylaws ... and the doing of all other acts incident to the legal status of stockholders. Nor will a duplication of some or all of the directors or executive officers be fatal. Id. at 61-62, 118 S.Ct. 1876 (quotation marks and citations omitted). Though acknowledging that respect for corporate *241 distinctions when the subsidiary is a polluter has been criticized, the Supreme Court observed that “nothing in CERCLA purports to reject this bedrock principle, and against the venerable common-law backdrop, the congressional silence is audible.” Id. at 62, 118 S.Ct. 1876. Thus, the Court noted that “when (but only when) the corporate veil may be pierced, may a parent corporation be charged with derivative CERCLA liability for its subsidiary’s actions.” Id. at 63-64, 118 S.Ct. 1876 (quotation marks omitted). Nevertheless, the Supreme Court also recognized that a parent corporation could be held directly liable under CERCLA for its own actions in operating a facility owned by its subsidiary. See id. at 65,118 S.Ct. 1876. For under CERCLA’s plain language, anyone who operates a polluting facility may be liable directly for that pollution, regardless of whether that person is “the facility’s owner, the owner’s parent corporation or business partner, or even a saboteur who sneaks into the facility at night to discharge its poisons out of malice.” Id. As the Supreme Court acknowledged, the difficulty comes in defining actions sufficient to constitute direct parental operation. Under CERCLA, the Court said, an operator is simply someone who “directs the workings of, manages, or conducts the affairs of a facility.” Id. at 66, 118 S.Ct. 1876. However, to “sharpen the definition” still further, the Court held that the “operator must manage, direct, or conduct the operations specifically related to pollution, that is, operations having to do with the leakage or disposal of hazardous waste, or decisions about compliance with environmental regulations.” Id. at 66-67, 118 S.Ct. 1876 (emphasis added); see also Miami-Dade County v. United States, 345 F.Supp.2d 1319, 1342 (S.D.Fla.2004) (“[Tjhis Court must examine the actions of the United States to determine whether such actions constitute direction or management of operations as specifically related to pollution.”). In resolving whether a parent corporation manages, directs, and conducts the operations of a facility that relate to pollution, courts should focus on the parent’s interaction with the subsidiary’s facility, and not on the relationship between the two corporations. As the Supreme Court put it, “[t]he question is not whether the parent operates the subsidiary, but rather whether it operates the facility, and that operation is evidenced by participation in the activities of the facility, not the subsidiary.” Bestfoods, 524 U.S. at 68, 118 S.Ct. 1876; see also id. at 71, 118 S.Ct. 1876 (“[T]he verb ‘to operate’ ... obviously mean[s] something more than mere mechanical activation of pumps and valves, and must be read to contemplate ‘operation’ as including the exercise of discretion over the facility’s activities.”). Thus, control of the subsidiary, if it is extensive enough, may give rise to indirect liability via piercing of the corporate veil, but it does not give rise to direct liability as an operator under CERCLA. The Supreme Court therefore distinguished between direct operator liability and indirect derivative liability. See Bestfoods, 524 U.S. at 70, 118 S.Ct. 1876; see also Basic Mgmt., Inc. v. United States, 569 F.Supp.2d 1106, 1117 (D.Nev.2008). The Court particularly faulted the district court in Bestfoods for blurring that distinction by focusing on the relationship between the two corporations and the fact that certain individuals who held offices with the parent corporation also held offices with the subsidiary and made decisions at the subsidiary. That fact, the Court emphasized several times, was entirely appropriate and could not aloné “serve to expose the parent corporation to liability for its subsidiary’s acts.” Bestfoods, 524 U.S. at 69, 118 S.Ct. 1876 *242 (“[C]ourts generally presume that the directors are wearing their subsidiary hats and not their parent hats when acting for the subsidiary (quotation marks and citation omitted); id. at 70, 118 S.Ct. 1876 (“[I]f the evidence of common corporate personnel acting at management and directorial levels were enough to support a finding of a parent corporation’s direct operator liability under CERCLA, then the possibility of resort to veil piercing ... would be academic.”); id. (“[A] participation-and-control test looking to the parent’s supervision over the subsidiary, especially one that assumes that dual officers always act on behalf of the parent, cannot be used to identify operation of a facility resulting in direct parental liability.”). For, as the Supreme Court recognized, “directors and officers holding positions with a parent and its subsidiary can and do change hats to represent the two corporations separately despite their common ownership.” Id. at 69, 118 S.Ct. 1876 (quotation marks omitted); see, e.g., Atlanta Gas, 463 F.3d at 1205 (“[Ajctivities of a parent with respect to its subsidiary consistent with corporate norms should not give rise to liability under CERCLA.”); Sehiavone v. Pearce, 77 F.Supp.2d 284, 291 (D.Conn.1999) (“The evidence presented ... is consistent with the traditional and typical relationship between a parent and subsidiary [and] is insufficient to establish that Union Camp managed, directed or conducted operations specifically related to the pollution at the North Haven plant.”). Instead, the Supreme Court held, lower courts must look to the parent corporation’s exercise of control over activities at the facility that relate to pollution. The Supreme Court identified three circumstances in which a parent might be held liable for the direct operation of its subsidiary’s facilities: * first, “when the parent operates the facility in the stead of its subsidiary or alongside the subsidiary in some sort of a joint venture”; * second, when a dual officer or director departs so far from the norms of parental influence; and * third, when “an agent of the parent with no hat to wear but the parent’s hat” manages or directs activities at the facility. Bestfoods, 524 U.S. at 71, 118 S.Ct. 1876; see, e.g., Atlanta Gas, 463 F.3d at 1205 n. 6; Consolidated Edison, 310 F.Supp.2d at 604-05; BP Amoco Chem. Co. v. Sun Oil Co., 166 F.Supp.2d 984, 990 (D.Del.2001). Recognizing that identifying such circumstances requires difficult line-drawing, the Supreme Court returned to the usual norms of corporate behavior, which are “crucial reference points” and remain undisturbed by any CERCLA provision. Bestfoods, 524 U.S. at 71, 118 S.Ct. 1876; see Consolidated Edison, 153 Fed.Appx. at 751; New York v. National Serv. Indus., Inc., 460 F.3d 201, 208 (2d Cir.2006) (noting that in Bestfoods the Supreme Court looked to general hornbook law on corporate liability). “Activities that involve the facility but which are consistent with the parent’s investor status, such as monitoring of the subsidiary’s performance, supervision of the subsidiary’s finance and capital budget decisions, and articulation of general policies and procedures, should not give rise to direct liability.” Bestfoods, 524 U.S. at 72, 118 S.Ct. 1876 (quotation marks omitted). According to the Court, the critical question in any case is, therefore, “whether, in degree and detail, actions directed to the facility by the agent of the parent alone are eccentric under accepted norms of parental oversight of a subsidiary’s facility.” Id. B. In the case before this Court, Plaintiffs make no claim under the second and third *243 scenarios of parental control hypothesized by the Supreme Court in Bestfoods. Instead, this case is about the first scenario — more specifically, whether UGI operated the Norwalk facility in the stead of its subsidiaries, and whether UGI operated the remaining eight facilities alongside CL&P in “some sort of a joint venture.” The contours of Bestfoods’s “some sort of a joint venture” standard — which was borrowed from the Sixth Circuit — is not entirely clear and certainly has not been clarified by the Second Circuit. Plaintiffs contend that this is a broad and flexible standard indicated by a joint undertaking featuring control over the facility. See Pis.’ Posh-Trial Brief [doc. # 133] at 8-9. They point out that under Connecticut law, a joint venture is a special combination of two or more persons who combine their property, money, and skill to seek a profit jointly in a single business enterprise without any actual partnership or corporate designation. See id. UGI, by contrast, argues that at a minimum, this standard should not flow from normal parental oversight of the subsidiary, but rather should require actual parental operation of the facility relating to pollution and not merely control of the subsidiary itself. See Def.’s Post-Trial Brief [doc. # 134] at 12-13. New cases have discussed at length the “some sort of a joint venture” language in Bestfoods. Some courts have required the parent and subsidiary to enter into a joint venture agreement of" some kind. Explaining this requirement, the Delaware District Court stated that “[i]f alleging the existence of an undocumented joint venture were sufficient to state a claim for direct operator liability of a parent corporation under CERCLA, the requirement that a plaintiff pierce the corporate veil to give rise to indirect liability of a parent corporation under CERCLA would be rendered useless.” BP Amoco Chem. Co., 166 F.Supp.2d at 994. Indeed, on remand from the Supreme Court in Bestfoods, the district court refused to find operator liability on a joint venture theory because, in part, there was no joint venture agreement between the parent and subsidiary. See Bestfoods v. Aerojet-General Corp., 173 F.Supp.2d 729, 754 (W.D.Mch.2001). Other courts have looked to whether the plaintiff has established the typical badges of a joint venture under common law, such as intént to form a joint venture as evidenced by an agreement, the sharing of profits and losses, and joint or shared control of the facility. See Miami-Dade, 345 F.Supp.2d at 1351. Still other courts have looked to whether the parent’s actions directed toward the facility (and specifically related to pollution) were eccentric under usual corporate norms. See, e.g., Basic Mgmt., 569 F.Supp.2d at 1116. In MiamiDade, the court also held that the absence of a document reflecting the parties’ relationship placed a particularly heavy burr den on the plaintiff to demonstrate the existence of a joint venture, reasoning that business relationships such as a joint venture are not “entered into in a casual manner.” 345 F.Supp.2d at 1351. While it remains unclear precisely what the Supreme Court had in mind in crafting the phrase “some sort of a joint venture,” the Court is satisfied that Bestfoods did not intend to require the existence of all the formal requisites of a joint venture under the common law. Justice Souter, who wrote Bestfoods, chose his words carefully and it can be no accident that he chose the locution “some sort of a joint venture,” as opposed to a “joint venture.” Therefore, the Court concludes that this standard is somewhat more flexible than the common law definition of a joint venture. On the other hand, it is apparent from Bestfoods that the “some sort of a joint venture” standard cannot be used to eviscerate the portions of the opinion that went to great lengths to emphasize that the usual, or non-eccentric, relationship *244 between a parent corporation and its subsidiary — including budget, bylaw and ordinary investor control, overlapping officers and directors, and articulation of general policies and practices — is insufficient to impose direct operator liability. The reality is that parent corporations will ordinarily be involved in the affairs of their subsidiaries and that involvement will — at a minimum — indirectly affect the subsidiary’s facilities. See Lucia Silecchia, Pinning the Blame & Piercing the Veil in the Mists of Metaphor: The Supreme Court’s New Standards for CERCLA Liability of Parent Companies and a Proposal for Legislative Reform, 67 Fordham L.Rev. 116,177 (1998). Yet, that type of ordinary parental involvement with the subsidiary cannot be a basis for imposing direct operator liability under Bestfoods. See Commander Oil Co. v. Bario Equip. Co., 215 F.3d 321, 329 (2d Cir.2000) (refusing to conflate the concepts of “owner” and “operator”). Nor can the “some sort of a joint venture” standard be used to undermine the Supreme Court’s requirement that operator liability must hinge on direct management and control over the aspects of the facility that involve pollution. Recognizing that divining precisely what the phrase “some sort of a joint venture” means is a hazardous task, this Court nonetheless believes that the Supreme Court appears to have envisioned some form of joint undertaking or effort by the subsidiary and the parent together to manage, direct, and conduct the affairs of the facility (specifically as it relates to pollution) that is markedly different from the usual and accepted norms for parental oversight over a subsidiary’s facility. See, e.g., Pinal Creek Group v. Newmont Mining Corp., 352 F.Supp.2d 1037, 1041-42 (D.Ariz.2005) (“[Ejvidence of the corporate norms is critical because the acts of direct operation that give rise to parental liability must necessarily be distinguished from the interference that stems from the normal relationship between parent and subsidiary.”) (quotation marks omitted); BP Amoco, 316 F.Supp.2d at 171 (rejecting operator liability because plaintiff failed to prove that the parent company’s employee’s activities were “eccentric under accepted norms of parental oversight of a subsidiary’s facility”). This interpretation of the Supreme Court’s construct comports with the balance of Bestfoods and is consistent with the manner in which the Sixth Circuit itself described the test embraced by the Supreme Court. See Cordova Chem. Co., 113 F.3d at 579 (“[A] parent might ..., as a sort of joint venturer, actually operate the facility alongside its subsidiary.”). Under that standard, therefore, a formal agreement describing the parent’s and subsidiary’s responsibilities with respect to the facility would be useful to an understanding of the respective roles of the parent and subsidiary. But it is not required if the facts otherwise show parental operation of the facility alongside the subsidiary in a manner that is eccentric or dramatically inconsistent with accepted corporate norms for parental control of a subsidiary’s facility. See Marsh v. Rosenbloom, 499 F.3d 165, 182 (2d Cir.2007) (“Bestfoods ... clearly admonishes courts to refrain from creating CERCLA-speeific rules in the face of applicable long-standing common law principles.”). With that standard of “some sort of a joint venture” in mind, the Court now turns to applying Bestfoods to the facts of this case. III. While nodding to the facts relating to individual facilities, Plaintiffs have focused most of their attention on what they claim is the overall control by UGI of its subsidiaries, including CL&P. Their approach is understandable in light of the fact, discussed below, that there is very little, if any, evidence that UGI managed, controlled, and directed operations at in *245 dividual MGPs in Connecticut. Plaintiffs contend that inquiry into the general relationship between UGI and its subsidiaries remains important, despite statements in Bestfoods to the contrary. The Court is skeptical of Plaintiffs’ claims, since Bestfoods emphatically directs lower courts’ attention to the relationship between the parent corporation and the subsidiary’s facility, not the relationship between the parent and the subsidiary itself. As the Supreme Court put it, “[t]he question is not whether the parent operates the subsidiary, but rather whether it operates the facility, and that operation is evidenced by participation in the activities of the facility, not the subsidiary.” Bestfoods, 524 U.S. at 68, 118 S.Ct. 1876. Nevertheless, since the Supreme Court also indicated that eccentricity or adherence to ordinary corporate norms remains a touchstone of the Bestfoods operator inquiry, the Court begins its discussion with an examination of the overall relationship between UGI and CL&P, as well as the subsidiaries that owned the Nor-walk facility before CL&P. 6 The Court concludes, and finds as a fact, that there was nothing eccentric, or contrary to ordinary corporate norms as recognized in Bestfoods, in the relationship between UGI, the parent, and its subsidiary CL&P, or the subsidiaries that owned the Norwalk MGP before CL&P. A. The Court will begin with the Norwalk MGP, which has a relatively complex ownership history. See Ex. 1095 (summarizing ownership history of Norwalk MGP). UGI began its involvement with the Norwalk MGP in 1900, when UGI acquired a majority of the voting stock of the Connecticut Railway and Lighting Company (“CR & L”), 7 which owned the Norwalk Gas Light Company, which, in turn, owned the Nor-walk MGP. See Ex. 333, 1111, 1593. In 1906, CR & L leased the Norwalk property to the Consolidated Railway Company (“Consolidated”). See Ex. 1006 at 3610. Thereafter, in 1910, the New York, New Haven & Hartford Railroad (“NYNHHR”), a successor in interest to Consolidated, assigned to the Housatonic Power Company “all of the franchises demised to ... Consolidated ... by [CR & L] by [the December 19, 1906 sublease].” Ex. 2233 at 51883. The Norwalk MGP was included in the properties assigned to Housatonic in 1910. In 1911, NYNHHR and Housatonic subleased the Norwalk MGP to United Electric Light & Water Company (“United”), once again a company in which UGI had no ownership interest. See Ex. 1006. In late 1916, UGI acquired an interest in United. On August 9, 1917, the Syndicate, a subsidiary of UGI, acquired an interest in several companies, including United and Housatonic. See id. Thereafter, the property and franchises of all of the companies conveyed to the Syndicate were transferred to CL&P’s predecessor, RRPC. See id. From 1917 to 1941, the Norwalk Gas Light Company was subleased and operated by CL&P. There is little question that the Norwalk MGP was not operated by UGI during the period 1906-1917, and Plaintiffs’ counsel effectively acknowledged as much at oral argument. From 1906 to 1917, the Nor-walk MGP was leased or subleased to corporations over which UGI had no ownership interest, and Plaintiffs concede that the Norwalk MGP is not discussed in any UGI Board or Committee minutes during *246 that time period. See Ex. 2035, 2036. Given the extensive discussion of other facilities in the UGI minutes from 1906-1917, the absence of any discussion of the Norwalk MGP is — in the words of Best-foods — “audible.” There simply is no evidence whatsoever that UGI operated the Norwalk MGP from 1906-1917, either alone or in concert with any subsidiary or other company, and the Court so finds. That leaves two further periods for discussion — (1) the period from 1900 to 1906, when UGI owned CR & L, which in turned owned the Norwalk Gas Light Company, which operated the Norwalk MGP; and (2) the period from 1917-1941 when CL&P operated the MGP. The Court will discuss the 1917-1941 time period in connection with its general discussion of CL&P. As for the 1900-1906 time period, Plaintiffs contend that the charters for UGI’s Works Committee, which operated from 1888 to 1904, and Executive Committee, which operated from 1904 to 1923, show an extraordinary level of involvement by UGI in the affairs of its subsidiaries and their facilities. The minutes of the Works Committee no longer exist, and Plaintiffs would like the Court to speculate about what the Works Committee discussed and how it controlled the Norwalk MGP based upon the descriptions of the Committee’s charge and newspaper articles describing, in general, UGI’s oversight of its properties. The Court declines to engage in the speculation Plaintiffs request. In the absence of any evidence of UGI’s management or control of the Norwalk MGP from 1900 to 1904, the Court cannot, and will not, invent it out of thin air. The Executive Committee minutes do contain some references to actions undertaken at Norwalk by the UGI Board or the Executive Committee between 1904 and 1906, but these actions all concern tracking requisitions and expenditures, certainly not operations or pollution at the site, and do not translate into UGI control of the Norwalk MGP during that time period. See Ex. 2036 (summarizing all actions of Executive Committee regarding Norwalk MGP). The Executive Committee’s actions are in accord with UGI President Thomas Dolan’s letter in 1900, in which he states that CR & L will conduct the business of the Norwalk Gas Light Company, but that CR & L’s “operation and management will be under the supervision of The United Gas Improvement Company.” Ex. 332. Such supervision, particularly with respect to requisitions and capital expenditures, is consistent with UGI’s role as a parent corporation. See Bestfoods, 524 U.S. at 72, 118 S.Ct. 1876 (holding that “supervision of the subsidiary’s finance and capital budget decisions” is consistent with parent status); see also Schiavone, 77 F.Supp.2d at 292 (review and approval of capital budget, including pollution control equipment, does not demonstrate direct management of operations related to pollution). There is no question that particularly in the early years of its existence, UGI kept a close watch on its subsidiaries and properties. However, the Court does not find UGI’s level of oversight over the Connecticut companies, or the Norwalk MGP in particular, eccentric or beyond the norms of corporate behavior. Unlike at the Waterbury North MGP, UGI did not have its own employee act as superintendent of the Norwalk MGP. Moreover, the local subsidiaries were run by their own Boards of Directors, and Plaintiffs have not submitted any evidence to show that those Directors did not manage, control, and operate the subsidiaries’ facilities, albeit under the general supervision of UGI. In 1902, President Dolan, described UGI’s business model as follows: The direct control and active management of the various companies in which your company is a shareholder are vest *247 ed with their respective Boards of Directors, and each company has a complete official organization of its own. But the Works Committee passes favorably upon all property improvement and extensions and contracts for supplies, before they are authorized ... We are therefore in constant and close touch with [the] management [of the subsidiaries] without in any sense making our company liable to the name of ‘Trust.’ ... [T]hrough such a system as outlined above there can be and is co-operation in improving details of management, and each company in which The United Gas Improvement Company is a shareholder has the benefit of the advice of that company’s experts ... Ex. 1017. That form of oversight — -which does exhibit close scrutiny by the parent— is not eccentric and merely reflects UGI’s status as a stockholder that wanted to watch where its money was invested. In Consolidated Edison, the district court considered largely the same record evidence regarding UGI’s oversight of certain New York subsidiaries and rightly rejected the argument that this evidence showed UGI’s management and control of facility operations. See Consolidated Edison, 310 F.Supp.2d at 608 (“UGI’s approval of WLC expenditures falls within the parameters of the parent-subsidiary relationship.”). The Second Circuit affirmed the district court in language that aptly applies to this case as well: Largely for the reasons identified by the district court, we conclude that Con Ed has pointed to no evidence that would allow a reasonable jury to conclude that UGI managed, directed, or conducted operations specifically related to pollution, that is, operations having to do with the leakage or disposal of hazardous waste, or decisions about compliance with environmental regulations. Its evidence, in general, consists of (1) overlapping officers and directors between parent and subsidiary, (2) close parental control of the subsidiaries’ expenditures, and (3) UGI’s enthusiasm for its subsidiaries to use its patented gas machinery. No evidence, however, rebuts the presumption that dual officers and directors can faithfully serve both parent and subsidiary. No evidence is inconsistent with UGI’s (or American Gas’s) role as an investor in its subsidiaries. In other words, no evidence would allow a reasonable jury to find that the conduct of UGI or American Gas meets the Best-foods standard.... 153 Fed.Appx. at 752. It is true that the Second Circuit is not bound by its summary orders. However, there is no reason (and Plaintiffs have not presented any) for this Court to ignore the appellate court’s determination that UGI did not, as a matter of law, operate its subsidiaries’ facilities when that determination was made on essentially the same evidence presented to this Court. As the district court noted in Atlanta Gas, “[c]ourts have made clear ... that review and approval of capital budgets or physical improvements, including those involving contamination-related equipment, are consistent with a traditional parent-subsidiary relationship.” 2005 WL 5660476, at * 11; see also Bestfoods, 524 U.S. at 72, 118 S.Ct. 1876. Therefore, the Court finds as a fact that UGI did not operate the Norwalk facility between 1900 and 1906, either in the stead of its subsidiaries or in some sort of a joint venture alongside its subsidiaries. B. In addition to the Norwalk MGP, a number of other companies that operated MGPs in Connecticut were acquired by the Service Company in the 1920s and then ultimately merged into CL&P. 8 The Court *248 finds as a fact that there was nothing eccentric or abnormal about UGI’s interaction or involvement with CL&P, or the MGPs owned by CL&P, from 1917 to 1941. For one, CL&P had its own officers. J. Henry Roraback, the “builder and developer of the CL&P system,” Ex. 1011, was never employed by UGI nor served as a UGI officer or director. By all accounts, Mr. Roraback was an independent spirit and forceful presence on the Connecticut scene. It is difficult in the extreme to believe that he was controlled by others, and certainly no evidence submitted to this Court suggests that. Tellingly, while CL&P’s corporate history pays extensive tribute to Mr. Roraback’s contributions to the company, UGI is hardly mentioned at all. 9 The Court finds that this was not an oversight by the company historians, but rather a recognition that CL&P employees did in fact manage, control, and operate its facilities and control its own destiny. Likewise, C.L. Campbell who served CL&P from its inception in 1917 as secretary-treasurer, later as vice-president and ultimately, as president (succeeding Mr. Roraback), was never employed by UGI; nor did he ever serve as a UGI officer or director. From the founding of CL&P, I.W. Day served as CL&P’s Vice President of Operations, the person in charge of management and operation of the MGPs. He, too, was never employed by UGI or served as an officer or director of UGI. Robert Knowlton stayed with the company long after UGI divested from CL&P, eventually becoming its president in 1948. CL&P also had its own Board, which consisted of between nine and sixteen directors, many of whom were pillars of Connecticut industry. These Board members owed duties to all of CL&P’s shareholders, not just to UGI, which owned only about sixty percent of the company’s stock (though at first, UGI’s interest was indirect through the Service Company and the Syndicate). At no time from CL&P’s inception until 1941 did the UGI directors constitute a majority of the CL&P Board. See Ex. 1013, 1014. In fact, the record shows that UGI’s nominees to the CL&P Board rarely attended CL&P Board meetings, which generally were held in Connecticut or New York, not Philadelphia, where UGI was based. Indeed, CL&P’s President, C.L. Campbell, in his history of the company, noted that UGI never voted its stock in its favor over the interests of CL&P: [S]o long as The United Gas Improvement Company had owned stock in this Company it had not voted its stock by the personal presence of any of its officers at a meeting, that it had always sent its proxy to be voted by officers of this Company or Directors not subject to its influence, that it had never voted its stock in favor of the election of a majority of Directors subject to its influence but had favored the election of a majority of Directors who favored a lo *249 cal control of the Company. He also stated that on no occasion had any contest over an election of Directors arisen and that since May 29, 1935, when questions of public interest had arisen — such as mergers — the approval of the Public Utilities Commission was obtained before The United Gas Improvement Company was allowed to vote its stock in this Company. Ex. 1744 at 6421; see also Ex. 1532. The Court finds that there is nothing in the record that casts doubt on Mr. Campbell’s assessment. The CL&P Board met regularly and kept detailed minutes. Those minutes reflect the fact that the CL&P Board itself was responsible for managing all aspects of its business. For example, the CL&P Board appointed committees, decided employment and payroll matters, appointed officers, hired employees, approved capital improvements and equipment acquisitions, issued loans, authorized studies, and managed securities among many other activities. See Ex. 1125-1204. The CL&P Executive Committee dealt with similar matters including approving pensions (Ex. 1288), loans (Ex. 1289, 1409), contributions to local organizations (Ex. 1318), and annual budgets (Ex. 1300). The CL&P Board also regularly declared dividends. Indeed, at one point, Mr. Roraback rebuffed an effort by UGI executives to convince CL&P to reduce its dividend. See Ex. 1380 at 2132. In 1920, the CL&P Executive Committee appointed the Sub-Committee on Management, consisting of Messrs. Roraback, Day, and Campbell, who met weekly with power to act upon matters the Committee felt needed to be settled promptly. See Ex. 1178. None of these men were employed by UGI. Very few of CL&P’s Board and Executive Committee actions required authorization or approval from UGI. CL&P’s Board of Directors and Executive Committee established their own policies and procedures, and adopted their own budgets. Thus, as detailed in UGI’s proposed findings of fact, the following actions were undertaken by CL&P without UGI’s prior recommendation or approval: (i) On October 26, 1917, the Board of Directors organized a salary dividend and welfare fund for its employees. See Ex. 1133 at 56421. There is no record of UGI authorizing the creation of this fund. (ii) On February 15, 1918, the Executive Committee appointed a Finance Committee to “consider and arrange for a plan for financing the requirements of the Company.” Personnel from UGI did not represent a majority of the members of this committee. See Ex. 1141 at 44. (iii) The Executive Committee also appointed a Special Committee on Power Development that was created to “consider the various reports of the engineers on the power developments of the Company.” Personnel from UGI did not represent a majority of this committee. See id. (iv) On January 17, 1919, a Committee was formed consisting of Mr. Day, Mr. Norris, and Mr. Campbell to “devise a more simplified method of approving requisitions to the end that the work of the Company could be carried along without delay, while at the same time proper consideration and review could be given to all expenditures.” Ex. 1157 at 86. Personnel from UGI made up only one-third of this committee. (v) On October 16, 1940, the Board of Directors established a policy for employees who were drafted or chose to enlist in the military services. See Ex. 1587 at 914. *250 (vi) On February 21, 1940, when C.L. Campbell recommended that CL&P’s investment in The Bristol Traction Company be sold, the Board collectively decided to pay retirement allowances to former Bristol employees, a portion of which was to be subsidized by the Welfare Fund. See Ex. 1585 at 831. These represent only a handful of the numerous examples in the record of CL&P’s independent actions. Indeed, the minutes from one 1927 UGI meeting state that its Management Committee should “free itself as much as possible from details in connection with the operations of the various companies and ... to put greater responsibility on the local managers of the companies.” Ex. 152 at 13309; see also Ex. 1651 (1927 UGI Circle) (stating that “[i]t is the policy of [UGI] to constantly to put more responsibility on the heads of the local companies.... The ultimate goal is supervision not direction”); Ex. 404 at 14356 (1925 UGI Minutes) (‘With the budget system in effect and basic principles outlined for the guidance of local managers, who are responsible for results of operations, the Management Committee ... will need less and less to consider details in connection with the supervision of the utility subsidiary companies.... ”); Ex. 118 at 30 (Letter from President Dolan to UGI stockholders) (“Your capital is invested in companies ... conducting their business so as to give satisfaction to the communities in which they perform their very important public functions, each self-sustaining.”). While Plaintiffs’ expert, Thomas Blake, did not believe that CL&P’s minutes reflected reality — as he put it, the CL&P Board did nothing whatever of “substance” — there is no basis in the record for his wild and unsupported supposition, and the Court rejects it as fanciful and lacking all credibility. If anything in this case is eccentric, it is the views of Mr. Blake. CL&P also reported regularly to the Connecticut Public Utility Commission (“PUC”), which had broad regulatory powers over public utilities, including CL&P. See, e.g., Ex. 2208 (CL&P’s 1921 Annual Report to PUC); Ex. 2207 (Letter from PUC to A.J. Campbell of CL&P). The PUC set rates and standards for public utility companies and conducted inspections of plants. See Def.’s Post-Trial Brief [doc. # 135] Ex. 1, 2. In addition, the PUC prepared an annual report that detailed “the general conduct and financial condition of all public service companies as ascertained by the commission from the returns of the companies and examinations by the commission.” Id. Ex. 2, § 28. This report was prepared after the PUC collected sworn annual report forms from each utility, which were “signed and sworn to by the president or vice-president and treasurer of the company, or by a majority of the trustees or receivers, making the same.” Id.; see also Ex. 2208. While Plaintiffs have denied that CL&P was lying to regulatory bodies about UGI control, that is the only way of reconciling Mr. Blake’s theory that CL&P’s Board did nothing of substance with the fact that CL&P officers were providing sworn statements to the PUC concerning CL&P’s activities. Once again, there is simply no evidence to support Mr. Blake’s opinions and the Court rejects them as contrary to the record. The record is full of other evidence that CL&P was an independent, fully-functioning company. CL&P had its own bank accounts. As noted previously, UGI did loan CL&P its initial seed-money (approximately $10 million), but that money was repaid by 1921. See Ex. 186. Thereafter, CL&P raised funds in the marketplace. CL&P also had a large workforce. By 1931, it totaled over 1,800 employees, more than 1,000 of whom had been employed by the company for more than five years. See Ex. 1495 at 35512. The number of *251 employees had risen to 2,300 by 1937. See Ex. 1006 at 3611. CL&P also had a sophisticated operational structure, including engineers, researchers, technicians, real estate personnel and others. CL&P had its own purchasing department, even though it also availed itself of the group purchases effectuated by UGI for all of its subsidiaries. CL&P’s Operating Superintendent’s Department included design engineers, construction engineers, production engineers, gas engineers, test engineers, and plant superintendents, among others. See, e.g., Ex. 1393 (job descriptions for Superintendent’s Department). District Managers reported to the Vice President of Operations, and numerous individuals— including, among others, the District Engineers and Gas Engineers who actually operated the MGPs — reported to the District Managers. A detailed CL&P organizational chart, prepared by Mr. Day, in 1929 is reproduced at Exhibit 1020. See also Ex. 1393. CL&P dealt with customers, contractors, suppliers, and municipalities on its own, and also sold tar on its own. See, e.g., Ex. 1395, 1520, 1521, 1525, 1528, 1606. [[Image here]] The evidence shows that the Connecticut MGPs — which produced gas 24 hours a day — were operated long before they were acquired by CL&P. There is no indication in the record that UGI, or CL&P for that matter, made wholesale changes in the employees who were operating the individual MGPs. Moreover, as Dr. Shifrin admitted, each of the MGPs was operated according to local conditions and “local people” were the ones making the decisions about which techniques were to be used. See Tr. Transcript at 184-86. And although UGI initially purchased an interest in public utilities so that it could promote the installation of the CWG process, *252 eight of the ten MGPs at issue in this case (including Waterbury North) had already installed the CWG process by the time UGI acquired an indirect interest in them. While Bestfoods properly recognizes that determining who turns the gauges and pumps on a day-to-day basis is not the critical issue for operator liability, there must be some record evidence that UGI, either separately or in a joint venture with CL&P, managed or conducted the operations at the MGPs. Yet, there is no such evidence. According to Dr. Shifrin and Mr. Blake, because certain CL&P employees, such as W.T. Jebb, had previously worked for UGI, they owed their loyalty to UGI and not to CL&P, and took orders from UGI (though there are no such orders in the documentary record), not CL&P. See Ex. 1004 (summarizing biographies of relevant personnel). In Mr. Jebb’s case, once he began to work for CL&P, he stayed there and retired from CL&P. See id. Also, Dr. Shifrin and Mr. Blake each rejected the notion that any employee who wore two hats — such as J.A. Perry, who was employed by UGI but also served as a consulting gas engineer for CL&P and served on CL&P’s Board of Directors — owed any obligation to the subsidiary when that individual was working on subsidiary business. The Court had this colloquy with Dr. Shifrin: THE COURT: So you really have the view that when anybody holds two hats, they actually only have one, and that’s the more superior entity, which would be UGI. That you cannot — that no human being can actually separate themselves and say, “Listen, when I’m acting for CL&P, I’m a CL&P person, I owe my allegiance to CL&P. When I’m acting in my capacity for UGI, I owe my allegiance to UGI, and I actually separate myself that way.” THE WITNESS: I agree. I believe that in most cases. I can’t think of an exception right now. They might be acting on behalf of the subsidiary but they know that what they’re doing is feeding the parent. THE COURT: I’m sure they do. I’m sure they do. But you just reject the notion that you could wear two hats and actually do different things wearing one hat than another, when you have a job description that says, you know, you are supposed to be in charge of everything. THE WITNESS: Yes, especially Perry’s job. Tr. Transcript at 246-47. This rank speculation by Dr. Shifrin and Mr. Blake about what went through Mr. Jebb’s and Mr. Perry’s minds when they were working on behalf of CL&P is contrary to the general presumption — reaffirmed in Bestfoods — that employees who work for the subsidiary owe their loyalty to the subsidiary. See Bestfoods, 524 U.S. at 69, 118 S.Ct. 1876; Raytheon Constrs., Inc. v. Asarco, Inc., 368 F.3d 1214, 1219 (10th Cir.2003); Datron, Inc. v. CRA Holdings, Inc., 42 F.Supp.2d 736, 748 (W.D.Mich.1999). There is simply no evidence to support these experts’ view. There is no occasion when these employees acted contrary to the interests of CL&P when they were wearing their CL&P hats. The Court therefore rejects the suppositions of Dr. Shifrin and Mr. Blake as unfounded and lacking all credibility. These experts also pointed to the fact that from time to time UGI assigned “cadet engineers” to work at various subsidiaries, including CL&P. However, the number of such instances is very small and certainly small in relation to the entire workforce at CL&P. These “cadet engineers”- — -who became employees of the subsidiary — could not and did not control CL&P, let alone the Connecticut MGPs, not least because they entered the subsidiary at the very lowest level of the corpo *253 rate hierarchy. See Ex. 1020. Dr. Shifrin also acknowledged that there is no evidence of UGI training any CL&P employees in Connecticut. See Tr. Transcript at 162. Moreover, CL&P hired its own “cadet engineers,” and regularly promoted and moved employees throughout the company. See Ex. 1985. CL&P set salaries for its employees, including its cadet engineers. In fact, in 1950, Mr. Campbell, who was by then Chairman of CL&P described the company’s commitment to its employees as follows: The management of [CL&P] always has felt that one of its prime duties was to develop and train men to become in due course officers of the Company; and in so doing we have paved the way for promotions within our own ranks; we have made it possible to promote younger men into positions of greater duty and responsibility, and have developed an esprit de corps which, though intangible, is nonetheless strong. Ex. 1012 at 10835. The Court finds as a fact that Mr. Campbell’s description is accurate. According to Plaintiffs, the 1927 Management Services contract is, in effect, a joint venture agreement between CL&P and UGI. See Ex. 1052. The Court disagrees. The range of services that UGI offered its subsidiaries, including CL&P, was quite broad (for example including legal services), but there is no evidence that CL&P availed itself of all, or even many, of those services. More to the point, the record is bereft of any indication that CL&P depended on UGI to operate or conduct the affairs of its facilities for CL&P, or even alongside CL&P. To the Court, this agreement — which was reflective of the prior conduct of business between CL&P and UGI — merely reflected the fact that UGI made its consulting services available to CL&P if it wished to use them,- and UGI would charge for its services. See Ex. 1051. More importantly, whether CL&P used the services offered was dependent on CL&P, not UGI. See, e.g., Ex. 1338 (UGI’s P.H. Gadsden reporting on a visit to Connecticut in which he discussed UGI’s services with Mr. Roraback, who told the UGI purchasing agent to come to Connecticut to try to convince the local purchasing agent of “the value of [UGI’s] services”). This Court therefore agrees with the Eleventh Circuit’s assessment • of a similar contract in Atlanta Gas: Our review of the record persuades us that the “management” contract between the'two entities (which apparently was executed in the later years of the relationship) reflects the nature of the relationship all along. The contract explains the relationship between the two entities as being based on St. Augustine’s ability to tap into UGI’s vast knowledge and resources. There is nothing in the contract that suggests that UGI was contracting to actively manage the St. Augustine facility itself; rather, the contract allows St. Augustine Gas to receive advice from UGI’s various skilled personnel, to have access to UGI’s network of professionals, and in general to benefit from access to UGI’s vast experience.... Throughout the contract, its contemplation was that UGI would provide recommendations to St. Augustine Gas. The only exception was that UGI would actually serve as St. Augustine Gas’s purchasing agent for significant purchases requested by St. Augustine.... In other words, in the language of Bestfoods, nothing in the contract suggests that UGI was to “manage, direct or conduct operations specifically related to pollution, that is, operations having to do with the leakage or disposal of hazardous waste.” Similarly, there is nothing in the record sug *254 gesting that, prior to the time of such a contract between the two entities, UGI was engaged in such activities. To the contrary, the record suggests a similar advisory role. Atlanta Gas, 463 F.3d at 1205-06. UGI’s willingness to play a consultant or advisory role does not transform UGI’s relationship with CL&P into a joint venture of any kind. Plaintiffs point to the fact that UGI hoped to conduct yearly audits of the facilities operated by its subsidiaries. However, that goal appears to have been aspirational only, as UGI had hundreds of subsidiaries operating many more facilities. There is no indication that UGI did, in fact, inspect each of CL&P’s facilities annually. In fact, Dr. Shifrin testified that there is no evidence that UGI ever visited many of the MGPs at issue in this case, let alone inspected every MGP operated by its subsidiaries every year. The only inspection reports (for Norwalk and Meriden Cooper Street) that are in the record appear to this Court to be inspections conducted by CL&P employees for CL&P’s own use, and the Court so finds. See Ex. 126 (1927 Norwalk inspection report); Ex. 1649 (1927 Meriden inspection report). There is no indication in the record that these inspection reports were provided to UGI. Moreover, a CL&P internal memo from 1935 reflects the fact that the company inspected its own facilities: “each Gas Plant Superintendent and each Gas Plant foreman as per the organization specifications, is charged with direct responsibility of inspection, operation and maintenance of gas holders, accessories and connections, under their jurisdiction.” Ex. 1029. In fact, on several occasions, UGI employees had to be told what was happening at CL&P’s MGPs, which is inconsistent with any claim that UGI was operating or managing the facilities, either separately or alongside CL&P. For example, in response to a request from UGI, Mr. Day wrote to tell UGI that purification equipment at the Meriden MGP had been dismantled years before. See Ex. 256. Moreover, in 1935, Mr. Jebb asked B.V. Pfeiffer at UGI how UGI inspected the holders at MGPs. See Ex. 167. Had Mr. Jebb been operating the facilities under directions or orders provided by UGI, he would not have needed to ask that question. In fact, Mr. Jebb reciprocated and sent Mr. Pfeiffer CL&P’s practices regarding holder inspections several months later. See Ex. 425. Again, such correspondence would not be needed if UGI and CL&P were jointly operating CL&P’s MGPs. 10 In fact, they were not. This Court has reviewed with care the various correspondence between the entities as well as the minutes of the Boards and Committees of UGI and CL&P. Throughout the minutes and correspondence, there are from time to time notations of “recommendations,” “advice,” “suggestions,” and “assistance” that UGI offered to CL&P. See, e.g., Ex. 234, 239. Plaintiffs’ experts contend that UGI did no such thing and that its “recommendations” were, in fact, edicts or directives that CL&P employees were required to follow. Once again, however, there is no evidence to support these hypotheses. There are numerous examples of CL&P rebuffing recommendations made by UGI, including rejecting the timing of audits, advertising suggestions, recommendations regarding the size of CL&P’s Board, and the amount of its dividend. See Ex. 1019 (summary of UGI recommendations rejected by CL&P). *255 There are also examples of CL&P challenging UGI’s billings or actions. See Ex. 1317, 1400; Tr. Transcript at 358-60. Given these circumstances, the Court sees no basis for construing the language of these documents differently from what was written — as recommendations and advice from a knowledgeable and sophisticated parent to a knowledgeable and sophisticated subsidiary. See also Atlanta Gas, 463 F.3d at 1206 n. 8. (“The record also indicates that [UGI’s committees] provided such advice to numerous similar facilities in which UGI had a similar equity interest. Significantly, [Plaintiff] has pointed to no evidence indicating that even such advice and consultation involved operations specifically related to the leakage or disposal of hazardous waste.”). As discussed previously, UGI also provided its subsidiaries with Engineering and Operating Notes on various aspects of MGP operation (though none included advice on tar, leaks or waste handling, see Tr. Transcript at 32) and conducted annual conferences where the engineers who worked at the subsidiaries nationwide could gather and share experiences. However, Dr. Shifrin acknowledged that these notes were mere advice to the subsidiaries, not directives. See Tr. Transcript at 69 (“No, they weren’t being directed, they were being informed of state of art technology.”). Moreover, he conceded that he does not know how the Operating Notes were used, if at all, by the Connecticut MGPs. See Tr. Transcript at 63. He also eventually, although belatedly, acknowledged CL&P’s own sophistication and aptly summed up what the record appears to show: THE WITNESS: I think the framework was if the plant was making gas relatively efficiently, which would be judged in terms of output per unit input of oil and coke, they [CL&P] were probably left to their own devices on the ground. If those efficiencies were out of whatever specification UGI felt was required or if the plant asked for help, itself, then Philadelphia would come in. THE COURT: Right. So during the relevant time period, were these particular plants being operated relatively efficiently or were they out of whack with whatever UGI demanded of them. THE WITNESS: I think they were generally efficient. THE COURT: So the local people were generally left up to their own devices with respect to those— THE WITNESS: Right. Tr. Transcript at 182-83. At closing arguments, CL&P seized on the FTC’s 1930 report authored by Lewis G. Pritchard. See Ex. 160. Edwin Harris authored another report about UGI and its subsidiaries for the FTC in 1933. See Ex. 157. The Court has reviewed these lengthy reports in detail and neither of them demonstrate that UGI was operating the Connecticut MGPs alongside CL&P in some sort of joint venture, or controlled CL&P. To the contrary, Mr. Pritchard assiduously uses the word “recommendation” and “advice” when he speaks of UGI’s assistance to the subsidiaries. And while he refers to periodic reports that UGI’s subsidiaries were required to make to UGI, those reports (none of which involved CL&P) are consistent with UGI’s investor status. More importantly, and a comment that Plaintiffs prefer to ignore, Mr. Harris concluded that having examined UGI and its subsidiaries, “UGI does not dictate to the management of the Service Company [which in turn owned CL&P]. The Service Company’s operations were conducted as freely as though it were an unrelated company.” Ex. 1083 at 44222; see also id. at 44349 (“[T]he rate of return on [UGI’s] investment [in CL&P] has not been excessive.”). Neither Mr. Harris nor Mr. Pritchard had any incentive to pull their *256 punches when it came to UGI at this point in time. Mr. Harris’s conclusion is consistent with what the Court itself finds based on the documentary record. In short, Plaintiffs have failed to shoulder their burden of showing that there was anything eccentric or unusual about the corporate relationship between UGI and CL&P as it relates to the MGPs, let alone pollution at the MGPs. There were some overlapping Board members and officers, but that is not unusual or eccentric. UGI also provided CL&P with advice and recommendations from time to time when CL&P sought its parent’s advice. As Dr. Shifrin acknowledged, UGI had considerable information about and experience with the operation of MGPs and it was only natural therefore, that CL&P would want to tap into that resource from time to time. But no evidence shows that UGI and CL&P were operating the MGPs as a joint venture or joint undertaking of any kind. Instead, the relationship between the companies was the typical parent-subsidiary relationship and fully consistent with UGI’s investor status. UGI may have set general policies and approved budgets that CL&P recommended for UGI approval. See Bestfoods, 524 U.S. at 62, 118 S.Ct. 1876 (articulating policies and procedures and approving budgets is insufficient to establish operator liability). But the Court finds as a fact that it was CL&P (and CL&P alone) that had responsibility for creating and implementing those budgets and managing and directing operations of its facilities, including the MGPs. See, e.g., Atlanta Gas, 463 F.3d at 1205; Consolidated Edison, 153 Fed.Appx. at 752; Basic Mgmt., 569 F.Supp.2d at 1116; Pinal Creek Group, 352 F.Supp.2d at 1041-42; Miami-Dade County, 345 F.Supp.2d at 1351; Consolidated Edison, 310 F.Supp.2d at 604-05; Aerojet, 173 F.Supp.2d at 754; BP Amoco, 166 F.Supp.2d at 990; Atlanta Gas, 2005 WL 5660476, at * 11; Schiavone, 77 F.Supp.2d at 291. Arrayed against the substantial body of case law just cited, Plaintiffs lean heavily upon United States v. Kayser-Roth Corp., 103 F.Supp.2d 74 (D.R.I.2000), aff'd, 272 F.3d 89 (1st Cir.2001). That case arose in an odd procedural context, one in which one district judge had to consider a motion to vacate a decades old declaratory judgment issued by another judge. By contrast to the evidence presented in this case, in Kayser-Roth the facts established that Kayser-Roth essentially was in charge of practically all of the facility’s decisions, “including those involving environmental concerns.” 272 F.3d at 102; see also Kayser-Roth, 103 F.Supp.2d at 81. Kayser-Roth made the ultimate decision to replace the facility’s traditional soap process with one that used the contaminating agent, TCE. See Kayser-Roth, 272 F.3d at 103; Kayser-Roth, 103 F.Supp.2d at 81. Kayser-Roth issued a directive requiring its subsidiary to notify the KayserRoth legal department of any correspondence regarding environmental matters, see Kayser-Roth, 272 F.3d at 102, and made the final decision on settlement when the United States sued the subsidiary for illegal waste water discharge. See Kayser-Roth, 103 F.Supp.2d at 81. Based on such findings, the original judge expressly found that “Kayser-Roth directed ... activities with respect to environmental matters, in general, and operation of the facility utilizing TCE, in particular.” KayserRoth, 272 F.3d at 103. There is no similar evidence here, and therefore, Kayser-Roth provides no guidance for this case. C. Having found that there was nothing eccentric or unusual about the corporate relationship between UGI and CL&P generally, the Court now turns to discussing *257 Plaintiffs’ evidence of alleged control by UGI of operations at the particular MGPs. The Court finds as a fact that Plaintiffs have failed to prove by a preponderance of the evidence that UGI and CL&P conducted operations at the MGPs in some sort of joint venture or joint undertaking, and that UGI’s involvement with the individual MGPs was consistent with its role as an investor and shareholder of, and consultant to, CL&P. For certain of the facilities, Plaintiffs have presented essentially no evidence that could be possibly construed to show UGI operation and direction of the MGPs. The Court will begin with those facilities. Rockville. The Rockville MGP produced coal gas from 1863 to 1889 and CWG from 1889 to 1929 under the auspices of the Rockville Gas Light Company and the Rockville-Willimantic Lighting Company. See Ex. 195. The Service Company acquired the Rockville-Willimantic Lighting Company in June 1930 and the company was merged into CL&P in October 1935. See Ex. 1369. The facility became inactive in 1930. See Ex. 195. Dr. Shifrin lists three “examples” of UGI’s control of the Rockville facility. Dr. Shifrin explained that his “examples,” are the best evidence he could find of UGI control. Tr. Transcript at 79. Those examples are as follows: * CL&P notified UGI of changes to its construction budget in 1931. * CL&P inspected Gas Holder No. 3 in 1934. * Gas Holders No. 1 and 2 were “likely demolished” and removed between 1929 and 1934. 11 Ex. 1 at 71. Mr. Blake lists another instance of UGI control, which is a 1932 letter from Mr. Jebb of CL&P to the manager of the Rockville facility indicating that UGI wanted its representative to be present before any purifying box in the system was changed. Ex. 2 at 55. None of these examples shows that UGI and CL&P were operating the Rockville MGP in a sort of joint venture or joint undertaking. Notification of a capital budget, and even approval of it by UGI, is not eccentric and cannot be a basis for imposing operator liability. See Bestfoods, 524 U.S. at 72, 118 S.Ct. 1876. Furthermore, the fact that CL&P inspected the holders at its Rockville facility and removed the holders cannot show UGI involvement, let alone control, over the aspects of the facility that causes pollution. CL&P conducted those inspections and operations, and there is no evidence of UGI involvement. It is true that UGI asked to be present when the purifying boxes were changed, but that was because, as Dr. Shifrin acknowledged, a UGI facility in Camden had an explosion when a purifying box was changed. See Tr. Transcript at 192. UGI’s interest in promoting safety at its subsidiaries’ facilities cannot be transformed into management and control of the facilities. See Datron, 42 F.Supp.2d at 748 (“[Cjonducting sporadic safety inspections are the kind of activities which are consistent with the parent’s investor status.”) (citation and quotation marks omitted). Indeed, the correspondence makes it clear that UGI had no idea when purifying boxes would be changed at CL&P’s facilities, a fact that puts the lie to Plaintiffs’ claim that UGI was in control of the facilities. The documentary record that does exist regarding Rockville shows it was operated by the managers on site, with oversight by Messrs. Jebb and Day of CL&P. See, e.g., Ex. 1483. *258 In sum, Plaintiffs have failed to show by a preponderance of the evidence that UGI and CL&P operated the Rockville MGP in some sort of joint venture or joint undertaking or that UGI operated Rockville in the stead of CL&P. Accordingly, UGI has no operator liability regarding the Rock-ville MGP. Middletown. Gas production began at this facility in 1853. CWG production began in 1886 and ended in 1928. See Ex. 304 at 74770. The Service Company (for which UGI had grandparent ownership) acquired the Middletown Gas Light Company in 1927 and the company was merged into CL&P the same year. See Ex. 308 at 13144; 1090. CL&P included the Middle-town facility in its system, and Middletown had its own manager, who reported to Mr. Day of CL&P. See Ex. 1372. Dr. Shifrin points to two examples of UGI control of this facility: * UGI approved a contract for gas oil purchase in 1927. * UGI Contracting Company contracted in 1927 to build a high pressure distribution main connecting Middletown and Meriden. Ex. 1 at 80. Mr. Blake offers an additional item — Mr. Pfeiffer of UGI wrote to a UGI engineer suggesting that he might want to send one or two men to “assist” CL&P in turning on the high pressure distribution main. See Ex. 2 at 54. For many of the same reasons identified with respect to Rockville, the Court fails to understand how these examples show that CL&P and UGI were operating the Middletown MGP in some sort of joint venture or joint undertaking. It is undisputed that UGI purchased feedstocks for its subsidiaries in order to get a better price than the subsidiaries could get on their own. As Dr. Shifrin acknowledged, bulk purchasing made sense and benefitted the subsidiaries. That activity, which UGI carried on in Philadelphia, does not show UGI control over the aspects of the Middletown facility that generate pollution. Nor does the other “evidence” point in the direction of operator liability. UGI Contracting Company was a subsidiary of UGI, and CL&P had a contractual relationship with that company to construct the mains. See Ex. 307 at 13698. CL&P’s contract with UGI Contracting Company was just one of many contracts it had with companies to perform services at that site. See Ex. 1445. Contracting with a subsidiary of UGI for certain construction services cannot possibly demonstrate UGI operation, management and control of the Middletown MGP. Moreover, the decision to contract with the UGI Contracting Company to build the main between Meriden and Middletown appears to have been made before UGI acquired an indirect interest in the company, see Ex. 307 at 13698, which Dr. Shifrin admitted would, if true, preclude this contract from being an example of UGI control. See Tr. Transcript at 211. The letter from Mr. Pfeiffer to Mr. Eiehengreen of UGI suggesting that he send someone to “assist” CL&P in “gassing of the mains and turning consumers’ services en route,” see Ex. 445, is nothing more than it appears — UGI offering assistance to CL&P — and the Court so finds. This letter certainly does not suggest that UGI had ultimate authority over the project or that anyone other than CL&P and its site managers were in charge at all times. In fact, the Court does not even know whether Mr. Eichengreen ended up sending engineers to Middletown; we only know that Mr. Pfeiffer made that suggestion. This is no different than the myriad of other examples where CL&P relied on UGI’s expertise and experience to increase the efficiency of its operations. In sum, Plaintiffs have failed to show by a preponderance of the evidence that UGI *259 and CL&P operated the Middletown MGP in some sort of joint venture or joint undertaking, or that UGI operated Middle-town in the stead of CL&P. Accordingly, UGI has no operator liability regarding the Middletown MGP. Winsted Gay Street. Coal gas was produced at this site from 1874 to 1888; CWG was produced there from 1889 to 1953. See Ex. 263 at 80269. The Winsted Gas Company operated this facility. The Service Company acquired the Winsted Gas Company in 1929, and it was merged into CL&P in 1931. See Ex. 182 at 3990; Ex. 185 at 34863. In the years following 1929, the Winsted operations were headed up by Winsted Gas Company District Manager A.C. Bristol and Winsted Manager J.E. Lynch, neither of whom were employed by UGI and both of whom were eventually employed by CL&P. See Ex. 1004. Winsted had a number of employees, none of whom were employed by UGI. See Ex. 1466,1584,1512,1527. Dr. Shifrin offers as examples of UGI “control” of this facility the following: * UGI asked to be informed of purifier box changeovers in 1932. * CL&P planned an installation of a tar separator in 1929 after nearby residents complained of tar discharges. * CL&P inspected the gas holders in 1935 and 1938. * CL&P inspected the works in 1935 and recommended elimination of steam discharge to gas holder No. 1, which caused tar emulsions. * Cruse-Kemper installed gas holder No. 3 in 1937. See Ex. 1 at 67. To these facts, Mr. Blake adds nothing. See Ex. 2 at 55. Needless to say, all of the activity described above that is attributable to CL&P, including activity that involves tar, has no relevance to UGI’s operator liability. The reason is simple. There is nothing in the record to show that UGI was even aware of CL&P’s actions regarding tar at this facility or CL&P’s own inspections. That UGI asked to be informed about purifier box changeovers was discussed earlier and that request certainly does not show that UGI and CL&P were operating the Winsted MGP as a joint venture of any kind. As UGI points out, the documentation shows conclusively that this facility was being managed and operated by CL&P, not UGI. For example, in a 1929 letter from W.T. Jebb, Assistant Operating Superintendent of CL&P, to A.C. Bristol, Mr. Jebb stated that in order to improve the present performance of the Winsted plant, he had arranged for the shipment of several carloads of Koppers coke, which he considered to be “more adaptable for water gas generating purposes than Troy coke.” Ex. 1456. He had also arranged for representatives from the Coke Company to go to Winsted the next week to spend as long as necessary to acquaint the Winsted operators with the proper use of Koppers coke, and “strongly recommend[ed]” that Bristol see to it that his “local operators acquaint themselves with everything that goes on.” Id. In a separate letter dated December 4, 1929, Mr. Jebb wrote to John Joy, Test Engineer of CL&P, that he had arranged for Mr. LaFleur, CL&P Chemist, to meet with representatives from the Connecticut Coke Company while in Winsted. See Ex. 1457. Similarly, in an October 20, 1936 letter, Mr. Jebb discussed a study conducted at the Putnam gas operations employing a mix of “soft coal and coke as a generator fuel,” as a way to find possible savings. Ex. 1563. Mr. Jebb explained that this mixture had been employed at the Winsted gas operations “very successfully with a saving of three or four cents per thousand of generator fuel.” Id. Mr. Jebb believed *260 similar savings could be available if this mix was employed at the Willimantic and Putnam gas operations. See id. In his October 9, 1929 inspection report memo to D.P. Allen of CL&P describing his inspection of the Winsted gas plant, L.G. Witmer of CL&P recommended that “the Thomas Recording Calorimeter now located in Bristol be transferred to the Winsted Plant for use on the make gas and to afford closer regulation of the BTU content of the gas going to the city”; that “an accurate record be kept of all boiler shutdowns and repairs made; that an accurate record be kept of all boiler fuel used”; that the facility should purchase a dump truck and portable conveyer; that “the steam meters now located in Middletown be transferred to Winsted and installed in each set”; and that a primer coat be applied to the 100,000 ct. ft. storage holder and repairing its leaks. See Ex. 1450. None of this activity involved UGI at all. In sum, Plaintiffs have failed to show by a preponderance of the evidence that UGI and CL&P operated the Winsted Gay Street MGP in some sort of joint venture or joint undertaking, or that UGI operated the facility in the stead of CL&P. Accordingly, UGI has no operator liability regarding the Winsted Gay Street MGP. Willimantic. Citizens Gas Light Company built and operated an MGP on this site from the late 1800s. In 1900, Willimantic Gas & Electric Company acquired the facility. See Ex. 2018 at 39721. The Rockville-Willimantic Lighting Company acquired the Willimantic Gas & Electric Company in 1914. See Ex. 1123. In turn, the Service Company acquired the Rockville-Willimantic Lighting Company in 1930, and in 1935 that company was merged into CL&P. See Ex. 1092 (summarizing the ownership history of the Willimantic MGP). The Willimantic MGP operated until 1953. See Ex. 2018 at 39721. In the years following 1930, the Willimantic MGP had its own set of operating superintendents and gas engineers, district engineers, district managers and other employees on site who managed and conducted operations there. See Ex. 1109, 1530, 1583. Dr. Shifrin’s examples of UGI control of the Willimantic MGP are as follows: * UGI instructed the MGP on purifier box operations and asked to be notified if boxes were changed in 1932. * CL&P notified UGI of changes to its construction budget in 1931. * CL&P inspected two gas holders at the site from 1934-1937. * CL&P inspected the works in 1938 and recommended elimination of steam discharge to the holder that caused tar emulsions. * United Engineers & Constructors installed unknown “gas plant equipment” in 1930. * In 1938, Kurt Richter, then a CL&P employee (although he had previously been employed by other UGI subsidiaries) filed a detailed report on gas plant operations at Willimantic, which included recommendations regarding tar production. Ex. 1 at 73. Mr. Blake does not add anything else. See Ex. 2 at 55. As noted previously, none of these examples detract from CL&P’s operations of this facility. Indeed, if anything these examples confirm CL&P’s operation of the Willimantic MGP. There is no evidence that UGI received Mr. Richter’s report or had any involvement with tar at Willimantic. See Ex. 287. That CL&P contracted with UE&C to install equipment at the site that CL&P desired to have on site cannot be used to show that UGI operated the Willimantic facility, either separately or in concert with CL&P. UE&C was a subsidiary of UGI and there is no claim of veil *261 piercing that would allow its involvement to be attributed to UGI. Moreover, as UGI points out, the documentary evidence, much of which UGI also relies on, points in precisely the opposite direction. For example, in a September 24, 1937 letter from Mr. Jebb, then CL&P gas superintendent, to H.D. Larrabee, CL&P manager at Willimantic, Mr. Jebb refers to CL&P’s annual inspections being part of CL&P’s “practice in connection with gas holder maintenance.” Ex. 1109 at 20718. He then summarizes the result of annual inspections of gas holders at Willimantic. See id. In an April 29, 1938 letter from Mr. Jebb to Mr. Larrabee, detailed aspects of the operations of the gas plant at Willimantic were discussed, especially in the attached report and recommendations from Mr. Richter, then a CL&P engineer. See Ex. 287. A November 6, 1939 letter from R.H. Knowlton, Vice-President of CL&P, to A.V.S. Lindley, Vice-President of CL&P, discusses taking the relief holder at Willimantic, “which is thought to be in poor repair,” out of service for examination and possible replacement. See Ex. 1580. And a gas sampling analysis was taken at Willimantic by CL&P Chemist, F.J. LaFleur in June of that year. See Ex. 1579. If anything, this documentary evidence confirms that CL&P was in charge of the operations at the Willimantic MGP. In sum, Plaintiffs have failed to show by a preponderance of the evidence that UGI and CL&P operated the Willimantic MGP in some sort of joint venture or joint undertaking, or that UGI operated Willimantic in the stead of CL&P. Accordingly, UGI has no operator liability regarding the Willimantic MGP. Waterbury South. The Waterbury Gas Light Company began coal gas production at Waterbury South in approximately 1918; CWG production began in 1928. See Ex. 211. The Service Company purchased the Waterbury Gas Light Company in 1929 and that company was merged into CL&P in 1931. See Ex. 174 at 61876. The CL&P Executive Committee appointed C.J. Allen to be Manager of the CL&P, Waterbury Division, which included the Waterbury South facility. See Ex. 1455 at 492. Numerous CL&P and Waterbury Gas Light Company employees worked at this site. See Ex. 1481 at 12423; 1650. CWG production ceased in 1935. See Ex. 212. Dr. Shifrin lists the following as examples of UGI’s control of this site: * UE&C installed two CWG sets and a holder in 1938. * UGI requested notification before purifier boxes were changed. * In 1934, UGI, CL&P, and Blaw-Know engineers cooperated in designing and installing a Blaw-Knox scrubber at Waterbury South. * In 1929, Edwin Hall provided a report to Mr. Perry on a visit to the facility regarding gas-making operations. * In 1935, a UGI engineer sent an in-site report back to UGI regarding purging of the coal gas plant at Waterbury South. Ex. 1 at 55-56. Mr Blake adds the 1932 report from a UGI engineer who witnessed the cleaning of one of the oxide boxes at Waterbury South. See Ex. 2 at 54-55. None of this evidence shows that UGI and CL&P were operating Waterbury South as some sort of joint venture. There certainly was cooperation between UGI and CL&P regarding aspects of the facility. But it is far too episodic to establish UGI’s joint control and management of the site. Over the course of eleven years, Plaintiffs proffer three examples of reports to UGI or cooperation in design. These are plants that were operating 24 hours of the day during those eleven years. *262 What exists shows that CL&P called upon UGI for assistance from time to time when CL&P decided it needed it, see Ex. 218 (reporting that CL&P teamed up with UGI and Blaw-Knox engineered to design a scrubber), and that UGI was overseeing major equipment purchases and plant changes, see Ex. 219 (discussing plant improvement needed to utilize heavy oil); Ex. 212 (discussing dismantling of CWG apparatus). Yet, it was CL&P that ran, managed and conducted operations at the site. See, e.g., Ex. 1553, 1573, 1576, 1581. As described previously, UE&C’s installation of equipment that CL&P ordered for the facility, as well as UGI’s request to be notified regarding the purifier boxes, do not show UGI control or management of the site. In sum, Plaintiffs have failed to show by a preponderance of the evidence that UGI and CL&P operated the Waterbury South MGP in some sort of joint venture or joint undertaking, or that UGI operated Waterbury South in the stead of CL&P. Accordingly, UGI has no operator liability regarding the Waterbury South MGP. Bristol. CWG production began at this site in 1905 and ended in 1928. See Ex. 298 at 74591. Bristol and Plainville Tramway Company owned and operated the site from 1905 to 1921, when it changed its name to Bristol and Plainville Electric Company. See Ex. 1369. In 1925, the Service Company purchased all of the stock of the Bristol and Plainville Electric Company, which in 1927 was merged into CL&P. See id. The properties of the Bristol and Plainville Electric Company were known as the “Bristol district” of CL&P. See Ex. 1367. The Bristol MGP had its own compliment of on-site employees who operated the facility, including a District Manager, G.E. Coekings. See Ex. 1369, 1466, 1546, 1582. Mr. Cookings, who never worked for UGI, was the District Manager of the Bristol and Plain-ville Tramway Company before joining CL&P, and also served on the Board of Directors of Bristol and Plainville Electric Company from 1925 forward. See Ex. 2031. The Bristol employees negotiated contracts with third-party contractors, arranged for inspections of the MGP by third-parties, and used the CL&P Research Department to investigate certain issues relating to the operation of the site. See Ex. 1343,1354,1453,1466. Dr. Shifrin provides the following examples of UGI control of the Bristol site: * J.A. Perry corresponded with G.E. Coekings regarding operation of water gas sets in 1926. * UGI approved a contract for one year’s supply of gas oil in 1928-29. * UGI Management Committee minutes reflect gas work inspections in 1928. * W.T. Jebb of CL&P sought the advice of J.A. Perry and B.V. Pfeiffer regarding Mr. Jebb’s proposal to abandon the site in 1929. Ex. 1 at 78. To this, Mr. Blake adds that a conference was held in Bristol in 1927 to discuss supplying gas to Terryville and UGI personnel attended the conference, that in 1927 UGI received a five-year gas forecast for Bristol, and in 1929, that Mr. Perry wrote an internal memo to another UGI employee about the possibility of using the Bristol site to supply peak load to CL&P. See Ex. 2 at 52-53. These examples do not show that UGI was operating the Bristol MGP alongside CL&P in some sort of joint venture. That UGI was purchasing oil for its subsidiaries does not demonstrate that it was operating the Bristol site. UGI did that for many subsidiaries, and Dr. Shifrin agreed that such group purchasing made sense and produced a lower price than would otherwise apply. The minutes that Dr. Shifrin cites for the inspections in 1928 states *263 merely that inspections were started at Bristol by W.T. Jebb, a CL&P employee, and R.P. Stacy. See Ex. 173 at 14416. There is no report on the inspections. At a later meeting, referred to by Mr. Blake, the minutes speak of forwarding a “joint report” (of both the Gas Works and Distribution inspection and the Electric Distribution System) to Mr. Day, at CL&P, with a recommendation that he (not UGI) consider combining certain work at the site to “produce economies.” Ex. 449 at 14444. This report and recommendation confirm that it was CL&P — not UGI — that was operating the MGP. Similarly, the correspondence between UGI and Mr. Cockings and Mr. Jebb all reflect CL&P seeking UGI’s advice on certain aspects of the Bristol site, including its closure, and receiving from UGI recommendations and advice. See Ex. 296 (Letter from UGI engineer to G.E. Cockings) (“I am sending you this Burlington report simply for your information to show you what can be done after getting down to the best operating conditions.”); Ex. 302 (Letter from Mr. Jebb to Mr. Perry) (“I would be pleased to receive your comments and advice on the above procedure.”); Ex. 303 (Letter from Mr. Pfeiffer to Mr. Jebb) (“I am in receipt of your letter of April 30th and wish to state that I do not recommend removing any equipment from the Bristol plant at the present time.”). It is apparent from this correspondence that CL&P was in charge of making decisions about the Bristol site and that CL&P was prudently seeking advice from UGI. Such examples of CL&P seeking UGI’s advice and recommendations do not transform the parent-subsidiary and consulting relationship into a joint venture of any kind. In sum, Plaintiffs have failed to show by a preponderance of the evidence that UGI and CL&P operated the Bristol MGP in some sort of joint venture or joint undertaking, or that UGI operated Bristol in the stead of CL&P. Accordingly, UGI has no operator liability regarding the Bristol MGP. Meriden Cooper Street. Coal gas production occurred at the Meriden Cooper Street beginning in approximately 1873 and CWG production began in 1906 under the ownership of the Meriden Gas Light Company. See Ex. 244 at 50458; Ex. 1369 at 72142. UGI acquired the stock of that company in 1924, see Ex. 1873, and it was merged into CL&P in 1926. See Ex. 1662 at 26346. The plant ceased producing manufactured gas in 1928. See Ex. 218. Meriden Cooper Street had its own large group of on-site employees who operated the facility, including an Operating Superintendent, Gas Engineer, District Engineer, and District Manager. See Ex. 1444, 1449, 1466, 1482, 1510, 1527, 1535, 1546. There is substantial correspondence within CL&P regarding operations at the Meriden Cooper Street site, including capacity, equipment, operations, and tar in the site’s holder. See Ex. 1350, 1391, 1427, 1444, 1465, 1466, 1541,1542. Dr. Shifrin cites the following as examples of UGI control of this site: *264 Ex. 1 at 63-64. To this, Mr. Blake adds that UGI received a five-year forecast for Meriden, and in 1929, an internal UGI memorandum discusses the possibility of operating Meriden to supply CL&P with peak load gas. See Ex. 2 at 51-52. *263 * UGI approved contracts for purchase of gas oil and coal in 1927-28. * UGI inspected the facility in 1927 and 1928. * A UGI engineer acted as a consultant for rebuilding of Gas Holder No. 2 in 1935 and reported on the purging of the gas plant. * UGI personnel assisted with installing a high pressure distribution main connecting Meriden and Middletown in 1927. * Repairs to Gas Holder No. 2 were “supervised” by UE&C in 1935. * UGI “approved the hiring” of a cadet engineer in 1927. *264 For reasons previously stated, the fact that UGI acted, in Dr. Shifrin’s words, as a consultant to CL&P, that UGI engaged in bulk purchasing for CL&P, or that CL&P entered into contracts with UE&C to perform construction work at the Meriden site do not show that UGI and CL&P were operating the site in some sort of joint venture. Nor does that fact that UGI assisted with a distribution main operating from Middletown to Meriden show UGI operation and management of the Meriden site. That UGI asked for five-year forecasts as to the facilities operated by its subsidiary, CL&P, is consistent with UGI’s investor status. The so-called “hiring” of cadet engineers for CL&P is not what Dr. Shifrin makes it out to be. The UGI minutes, which he cites, states that UGI surveyed its subsidiaries to determine what their needs for cadet engineers were and CL&P responded that its Meriden facility could utilize one cadet engineer. See Ex. 142 at 12949. The minute goes on to state that the young engineers would be interviewed by the subsidiaries. See id. That Meriden, which had a large compliment of its own employees, ended up hiring one UGI-trained cadet engineer in a single year hardly constitutes evidence that UGI and CL&P were jointly operating the Meriden facility. To the contrary, this evidence shows only that CL&P was in charge of its workforce and its facilities. In sum, Plaintiffs have failed to show by a preponderance of the evidence that UGI and CL&P operated the Meriden Cooper Street MGP in some sort of joint venture or joint undertaking, or that UGI operated Meriden Cooper Street in the stead of CL&P. Accordingly, UGI has no operator liability regarding the Meriden Cooper Street MGP. Putnam. CWG production began at the Putnam MGP in 1878 under the Putnam Gas Light Company, see Ex. 291, which was then purchased by the East Connecticut Power Company (“ECPC”) in 1922. See Ex. 293. UGI acquired ECPC in 1926. See Ex. 327 at 39789. ECPC was merged into CL&P in 1929. See Ex. 291. CWG production ceased and the majority of the equipment was dismantled in 1953. See id. There was a large operating staff at Putnam, including an Operating Superintendent, an Assistant Operating Superintendent, and a Gas Engineer. See Ex. 1114. Thomas G. Ralph was the CL&P District Manager for Putnam from 1930 to 1940. He was never employed by UGI. See Ex. 1004. Likewise, L.B. Richards served as Putnam Engineer from 1929 to 1937. See id. Although he had been employed by other companies in which UGI had an interest prior to his employment with CL&P, he had never been employed by UGI. There is ample correspondence in the record between CL&P and Putnam employees concerning proper use of equipment, plant efficiency, and inspections. See, e.g., Ex. 1109, 1466, 1563. Dr. Shifrin cites the following as examples of UGI control of this site: * In 1928, UGI approved a purchasing agent and other technical and engineering positions for Eastern Connecticut Power Co. * In 1928, UGI approved a six-day work week for Eastern Connecticut Power Co. * UGI approved a contract for a one-year supply of gas oil in 1928-29. * Management Committee minutes mention a Putnam plant inspection in 1928. *265 * UGI prepared instructions for changing purifier boxes in 1932, and requested notification prior to changing any purifier box. * In 1938, Kurt Richter, then a CL&P employee, filed a detailed report on gas plant operations at Putnam, which included recommendations regarding tar production. Ex. 1 at 75. Mr. Blake adds that Management Committee minutes note the forwarding of a “joint report” of the Putnam inspection in 1928 to I.W. Day at CL&P. He also points to a letter written by Mr. Jebb to CL&P Superintendent H.D. Larrabee concerning a new method of manufacturing gas at the Putnam MGP. Most of these examples have already been discussed. There is no evidence that UGI received Mr. Richter’s report and the fact that he had previously worked for other subsidiaries of UGI does not raise the presumption that he was acting on behalf of UGI when he prepared the report as an employee of CL&P; that UGI was purchasing oil for its subsidiaries does not demonstrate that it was operating the site; there is no evidence that UGI participated in the 1928 inspection that produced the “joint report.” See Ex. 449. A close read of the Management Committee minutes makes clear that the word “joint” refers to the fact that the inspection covered both the Putnam and Danielson Districts, not that UGI was somehow participating in the inspection. See id. As for the letter between Mr. Jebb and Mr. Larrabee, it is difficult to understand how Plaintiffs seek to use a letter between two employees of CL&P concerning a CL&P MGP as evidence of UGI control of the MGP. Only two of Mr. Shifrin’s examples of UGI control require further discussion— that UGI approved a purchasing agent for the Putnam Light Gas Company in 1928 and that UGI approved a six-day work week for Eastern Connecticut Power Company employees, also in 1928. Both of these examples occurred before CL&P owned or operated the Putnam MGP. In the first case, UGI Management Committee minutes reflect that J.A. Pearson, a vice-president of UGI, was approved as a purchasing agent for Eastern Connecticut Power Company and the Putnam Light Gas Company, a position he also held for CL&P. See Ex. 142. Mr. Pearson stated that this appointment was necessary because otherwise he had no authority to sign contracts on behalf of ECPC. See id. The Court has already made clear that the mere fact that UGI was making bulk purchases for its subsidiaries does not mean that it operated the MGPs in some sort of joint venture. Therefore, even if Mr. Pearson was only acting in his capacity as a vice-president of UGI, the fact that he was making purchases on behalf of the Putnam Light Gas Company would not indicate UGI control of the facility. The fact that Mr. Pearson became a purchasing agent for Eastern Connecticut Power Company and the Putnam Light Gas Company is even more problematic for Plaintiffs because at that point he had a ECPC hat to wear. There is no evidence in the record to suggest that Mr. Pearson was acting on UGI’s behalf when he was acting in his role as purchasing agent for ECPC. And, as the Court has reiterated multiple times, the presumption is that dual officers can and do wear different hats when working for the parent and when working for the subsidiary. The fact that UGI approved his appointment also does not rebut this presumption. There is nothing in the minutes that reflects ECPC’s position on Mr. Pearson’s appointment. We do not know whether ECPC made the initial decision to appoint Mr. Pearson, whether ECPC agreed with UGI’s recommendation on the matter, or *266 whether, as Plaintiffs’ experts no doubt believe, Mr. Pearson’s appointment was forced upon ECPC against its will. The Court will not invent such evidence in the absence of anything in the record. The Management Committee did approve a six-day work week for its subsidiaries in 1928 and ECPC is one of the subsidiaries cited in the January 24, 1929 minutes as “now on a six days per week working schedule.” See Ex. 297 at 014617. However, the minutes from the May 24, 1928 meeting show that ECPC was already on a six-day work week when the decision was made. See Ex. 2041 at 2188 (citing a report by J.A. Perry detailing companies that still had seven-day work week, not including ECPC). In other words, the Management Committee’s approval of a six-day work week had absolutely no effect on employees at the Putnam MGP, who were already working six days a week. Furthermore, how this evidence relating to work weeks shows UGI and CL&P joint control of operations with respect to pollution at the Putnam MGP is never explained by Plaintiffs. And for good reason. It is not evidence of joint management and direction of the MGP’s operations, nor is it at all related to pollution. In sum, Plaintiffs have failed to show by a preponderance of the evidence that UGI and CL&P operated the Putnam MGP in some sort of joint venture or joint undertaking, or that UGI operated the Putnam MGP in the stead of CL&P. Accordingly, UGI has no operator liability regarding the Putnam MGP. Norwalk. Having already given the history and ownership structure of the Norwalk MGP, the Court turns directly to the remaining 1917-1941 time period. There are, admittedly, more examples of UGI involvement with the Norwalk MGP than with many of the other MGPs, but the Court does not believe that this involvement amounts to “some sort of a joint venture” under Bestfoods. Dr. Shifrin and Mr. Blake give the following examples of UGI control of the Norwalk MGP during this time period: — 1922 — * UGI Contracting Company was contracted to remodel the Norwalk MGP. — 1925 — * UGI’s B.V. Pfeiffer wrote a letter to the District Superintendent of CL&P at the Norwalk MGP discussing the recommendation, details, and finances of installing a waterless storage holder and its advantages over a tank holder. ■1926 * Mr. Pfeiffer inspected the plant following complaints by the community of tar/oil in the Norwalk River and reported his findings to J.A. Perry. — 1927 — * Mr. Pfeiffer re-inspected the plant, again reporting his findings to Mr. Perry, and made several recommendations for dehydrating tar stored in the relief holder for sale and for better tar separator operation. * UGI conducted a detailed inspection of the Norwalk MGP. * J.A. Perry wrote to C.L. Campbell providing instructions for “the proper purging of old gas apparatus before undertaking repairs, and the proper purging of new gas apparatus before putting into use.” * J.E. King of CL&P responded that the tar separator “operation is now exactly in accordance with Mr. Pfeiffer’s wishes.” * UGI received a five-year forecast of the Norwalk MGP. * Mr. Campbell wrote to UGI requesting that an engineer from UGI review holder capacity at the Norwalk MGP. *267 1928 * Norwalk was still having problems operating its tar separator and requested help from Mr. Pfeiffer, who recommended a new tar separator. * UGI Contracting Company installed a gas holder. * UGI approved a contract for a one-year supply of gas oil in 1927-28. — 1929 — *An issue arose regarding disposal of effluent at Norwalk and CL&P requested guidance from UGI on improving the operations of tar separators. * A report was provided to UGI regarding the results of a recent inspection trip. The inspection and subsequent changes related to the use of heavy oil in place of gas oil, tar separators and extractors, tar conditioning and storage, and other specific gas-making procedures. * CL&P responded to UGI’s request regarding the “sale of water gas tar from Norwalk.” * United Engineers and Constructors installed equipment at MGP. * D.P. Allen of CL&P wrote to E.A. Harris of Norwalk with detailed suggestions for reducing the load on Nor-walk’s separators and filters. — 1932 — * UGI requested notification prior to opening any on-site purifier box. — 1936 — * Leaks were caulked in gas holder. Ex. 1 at 59-61; Ex. 2 at 32-33. 12 The first thing that is immediately apparent is that most of these examples occurred between 1926 and 1929. Of the examples outside that time period, the Court has dealt with the majority of them in relation to other plants. Actions taken by UGI Contracting Company and UE&C are not attributable to UGI; UGI’s request for five-year forecasts and for notification before removing any on-site purifier box are consistent with its role as a parent corporation; bulk purchasing for its subsidiaries does not establish UGI control. As for the leaks that were caulked in 1936, there is a reason Dr. Shifrin framed this example using the passive voice — the record does not indicate who caulked the leaks, whether UGI had any involvement in it, or even if UGI knew anything about it. See Ex. 242. Of the examples between 1926 to 1929, many of them involve a single event involving the failure of a tar separator that resulted in tar flowing into the Norwalk River. The first reference to this event is Mr. Pfeiffer’s report to Mr. Perry in 1926. In that report, Mr. Pfeiffer states that tar had been flowing into the Norwalk River for “a great many years,” Ex. 147, and makes a series of recommendations on how to solve the problem. At Mr. Pfeiffer’s follow-up inspection of the plant six months later, he makes further recommendations. See Ex. 145. The next correspondence between UGI and CL&P about the tar separator comes in 1929 when, in response to a request from CL&P, Mr. Pfeiffer writes a letter giving recommendations for improving the tar separator’s operation. Ex. 234. The Court finds several aspects of this correspondence revealing. First of all, while Mr. Pfeiffer does use the word “instruct” at one point, his letters are mostly framed as “recommendations” or “recommended” actions. Plaintiffs would like the Court to interpret UGI’s recommendations as directives, but as the Court has previously explained, the Court takes at face value the language in the documents, ab *268 sent some evidence that it should interpret it otherwise. In an effort to show that Mr. Pfeiffer’s “recommendations” were actually directives, Plaintiffs point to a 1927 letter from CL&P’s J.B. King to Alex Campbell, the General Superintendent of CL&P, which states that “the operation [of the Norwalk MGP] is now exactly in accordance with Mr. Pfeiffer’s wishes.” Ex. 143. Mr. Blake in particular believes that this is incontrovertible evidence that CL&P was implementing UGI’s directives, not following its recommendations. However, Mr. Blake fails to notice that Mr. King’s letter refers to Mr. Pfeiffer’s plan of action as a “recommendation.” And there is no evidence that Mr. King or Mr. Campbell reported back to Mr. Pfeiffer that his “wishes” had been fulfilled, as one would expect them to do if they had, in fact, received directives. This brings the Court to the most glaring problem with Plaintiffs’ theory of UGI control. As Dr. Shifrin admitted at trial, tar had been flowing into the Norwalk River for years before UGI knew about it. It was not until Mr. Briggs, the Norwalk District Superintendent, alerted UGI to the problem that UGI became of aware of it. See Ex. 147 (“In accordance with a letter dated August 28 instant, from Mr. C.D. Briggs, District Superintendent at Norwalk, Connecticut, to you, I visited the gas plant at Norwalk....”). It strikes the Court as completely implausible that an operator of a facility would need to be told about a major equipment failure and would not be told until years after the problem first appeared. Moreover, UGI waited six months before following up after the first inspection. The next correspondence about the tar separator occurs a full year later and once again it was initiated by CL&P. This is perhaps the strongest example of UGI involvement with pollution at any of the MGPs in question. In fact, Dr. Shifrin testified that other than Norwalk, there is no documentary evidence of any UGI involvement with tar at any of the other MGPs. See Tr. Transcript at 111-12. Given Bestfoods emphasis on control of the operations that produce pollution, that acknowledgment is significant. Yet, even in Plaintiffs’ best example, UGI’s involvement amounted to two visits and three letters in three years — hardly the “pervasive” control that Plaintiffs argue existed and certainly not “joint operation.” To be sure, UGI was concerned about a major equipment failure at one of its subsidiaries and CL&P drew upon the expertise and experience of UGI in order to solve the problem. But this, as the Court explained above, is insufficient to establish that UGI operated the facility. Plaintiffs point to several other letters from Mr. Pfeiffer that contain recommendations for the Norwalk MGP. See, e.g., Ex. 234, 386. From these letters and Mr. Pfeiffer’s involvement with the tar spill, they extrapolate that Mr. Pfeiffer was operating the Norwalk MGP on behalf of UGI from 1917 to 1941. However, as Dr. Shifrin admitted, Mr. Pfeiffer was physically at the Norwalk MGP for a total of approximately five days during that fourteen-year time period. See Tr. Transcript at 88. Although the Court understands that an operator does not need to be “turning the knobs,” as Dr. Shifrin put it, a few day trips over the course of a decade and a half is not sufficient to show UGI’s management and direction of the MGP. Furthermore, all of these letters detail recommendations, not directives, most of which were prompted by CL&P requests for advice or assistance. There are only a few other examples left and, again, none of them establish UGI control of the Norwalk MGP. The 1927 letter from J.A. Perry — written on CL&P stationary — to C.L. Campbell concerning *269 proper purging procedures is correspondence between two CL&P employees and cannot show UGI control. See Ex. 236; see also Ex. 387 (request from Mr. Perry for “weekly operating data”); Ex. 219 (report to Mr. Perry regarding recent inspection at Norwalk). Plaintiffs insist that Mr. Perry was also acting on behalf of UGI, but the Court rejects this argument as inconsistent with the presumption in Best-foods. The 1929 letter about separators and filters is also between two CL&P employees, neither of whom worked for UGI. See Ex. 392. In fact, these are just two examples of the ample correspondence between CL&P employees about various aspects of operating the Norwalk MGP. See Ex. 1205, 1330, 1451, 1536, 1555, 1561. This correspondence merely confirms that it was CL&P that was operating the Nor-walk MGP. In sum, Plaintiffs have failed to show by a preponderance of the evidence that UGI and CL&P operated the Norwalk MGP in some sort of joint venture or joint undertaking, or that UGI operated the Norwalk MGP in the stead of CL&P. Accordingly, UGI has no operator liability regarding the Norwalk MGP. IV. As stated above, the Court has concluded that there is no operator liability under CERCLA for the MGPs in question (with the exception of Waterbury North, which will be dealt with in the second phase of trial, if necessary). In the alternative, however, the Court finds that the cost recovery actions for two of the MGPs— Norwalk and Willimantic — are also barred by the statute of limitations. 13 The Court reaches these issues because they were presented during the first phase of the trial and in order to facilitate appellate review. A. Under CERCLA, there are two applicable statutes of limitations depending on whether the cost recovery action is characterized as a remedial action or a removal action. A cost recovery action under § 107(a) must be brought “for a removal action, within 3 years after completion of the removal action” and “for a remedial action, within 6 years after initiation of physical on-site construction of the remedial action....” 42 U.S.C. § 9613(g)(2). The statute of limitations is an affirmative defense and, thus, UGI has the burden to show that the cost recovery actions are time barred. See Chimblo v. Commissioner, 177 F.3d 119, 125 (2d Cir.1999). Plaintiffs filed suit in this case on September 1, 2006, meaning that for a remedial action to be timely, physical on-site construction must have been initiated no later than September 1, 2000. Likewise, a removal action completed after September 1, 2003 would be timely. Because of the different limitations periods and triggers, “whether an activity is a ‘removal action’ or a ‘remedial action’ under § 107(a) can be determinative of the timeliness of a claim.” OBG Tech. Servs., Inc. v. Northrop Grumman Space & Mission Sys. Corp., 503 F.Supp.2d 490, 524 (D.Conn.2007) (quoting Schaefer v. Town of Victor, 457 F.3d 188, 195-96 (2d Cir.2006)). Unfortunately, while CERCLA contains detailed definitions of remedial and removal actions, the overlapping definitions and the complexity of most CERCLA cleanups often make it difficult to determine whether an action should be *270 classified as a remedial action, a removal action, or both. The statute defines a removal action as: the cleanup or removal of released hazardous substances from the environment, such actions as may be necessary taken in the event of the threat of release of hazardous substances into the environment, such actions as may be necessary to monitor, assess, and evaluate the release or threat of release of hazardous substances, the disposal of removed material, or the taking of such other actions as may be necessary to prevent, minimize, or mitigate damage to the public health or welfare or to the environment, which may otherwise result from a release or threat of release. 42 U.S.C. § 9601(23). A remedial action, on the other hand, is defined as: those actions consistent with permanent remedy taken instead of or in addition to removal actions in the event of a release or threatened release of a hazardous substance into the environment, to prevent or minimize the release of hazardous substances so that they do not migrate to cause substantial danger to present or future public health or welfare or the environment. The term includes, but is not limited to, such actions at the location of the release as storage, confinement, perimeter protection using dikes, trenches, or ditches, clay cover, neutralization, cleanup of released hazardous substances and associated contaminated materials, recycling or reuse, diversion, destruction, segregation of reactive wastes, dredging or excavations, repair or replacement of leaking containers, collection of leachate and runoff, onsite treatment or incineration, provision of alternative water supplies, and any monitoring reasonably required to assure that such actions protect the public health and welfare and the environment. 42 U.S.C. § 9601(24). Distilled to its core, the key distinction between a remedial and a removal action is the purpose for which the action is undertaken. As a rule of thumb, though not an invariable guide, remedial actions are “long-term or permanent containment or disposal programs,” New York v. Shore Realty Corp., 759 F.2d 1032, 1040 (2d Cir.1985); see also Schaefer, 457 F.3d at 195; OBG Tech. Servs., 503 F.Supp.2d at 523-24, while removal actions are “typically short-term cleanup arrangements,” id. at 524 (quoting Shore Realty Corp., 759 F.2d at 1040). In general, “a removal action costs less [than a remedial action], takes less time, and is geared to address an immediate release or threat of release.” Colorado v. Sunoco, Inc., 337 F.3d 1233, 1240 (10th Cir.2003). The statute clearly contemplates that both a remedial action and a removal action can occur at the same facility. See 42 U.S.C. § 9601(24) (stating that remedial action includes measures “taken instead of or in addition to removal actions in the event of a release or threatened release of a hazardous substance into the environment”); 42 U.S.C. § 9613(g)(2) (“[I]f the remedial action is initiated within 3 years after the completion of the removal action, costs incurred in the removal action may be recovered in the cost recovery action.”). However, courts have generally held that there can be only one removal and one remedial action per facility, regardless of the number of phases in which the cleanup occurs. See, e.g., Sunoco, 337 F.3d at 1241; Kelley v. E.I. DuPont, 17 F.3d 836, 844 (6th Cir.1994). Moreover, and as all parties agree, a site with a single source of pollution is almost always considered one “facility” within the meaning of CERCLA and is generally not divisible absent ex *271 traordinary circumstances. See, e.g., Sierra Club v. Seaboard Farms, Inc., 387 F.3d 1167, 1170 (10th Cir.2004); Axel Johnson, Inc. v. Carroll Carolina Oil Co., 191 F.3d 409, 417 (4th Cir.1999); United States v. 150 Acres of Land, 204 F.3d 698, 709 (6th Cir.2000); New York v. Westwood-Squibb Pharm. Co., 138 F.Supp.2d 372, 376 (W.D.N.Y.2000); Niagara Mohawk Power Corp. v. CONRAIL, 291 F.Supp.2d 105, 125 (N.D.N.Y.2003). To make matters more complicated, there are some cases in which work on a site is neither a remedial nor a removal action, but rather constitutes “preliminary” or “interim” measures that do not trigger the statute of limitations. For instance, “evaluation, sampling, surveying and measuring” do not constitute “the initiation of on-site construction because these activities [do] not constitute ‘construction.’ ” Schaefer, 457 F.3d at 204 (quoting United States v. Findett Corp., 220 F.3d 842, 848 (8th Cir.2000)). Similarly, measures that are “inconsistent with the federal EPA’s remedial plan” do not start the statute of limitations on a remedial action. Schaefer, 457 F.3d 188 at 204 n. 20 (quoting GenCorp, Inc. v. Olin, Corp., 390 F.3d 433, 443-44 (6th Cir.2004)); 42 U.S.C. § 9601(24) (defining a remedial action as “those actions consistent tvith permanent remedy ”). Plaintiffs make two main arguments about why the statute of limitations has not run on the facilities at Norwalk and Willimantic. They argue that (a) the actions that occurred at Norwalk and Willimantic constituted removal actions, not remedial actions, and are not yet complete and (b) even if the work at both sites were considered remedial, UGI cannot prove that it was performed consistent with a “permanent remedy” as required by the statute. B. For the most part, the parties are in agreement about the facts that give rise to UGI’s statute of limitations defense. Their disagreement arises from how to characterize those facts. Thus, with respect to each MGP, the Court will make its findings of fact and then address the legal question of whether an action occurred at the facility and if so, whether to characterize it as a remedial or removal action. See OBG Tech. Servs., 503 F.Supp.2d at 524 (stating that whether an action is removal or remedial is a question of law, citing cases). Norwalk. Plaintiffs submitted a Site Remediation Documentation Report for the Norwalk MGP, dated August 7, 1995, that outlines the actions that occurred at the facility from September through November 1994. See Ex. 231. According to the report, approximately 9,000 cubic yards of contaminated soil and at least one small tank (or tar vat) were removed from the site “in advance of the construction of the new Yankee Gas Service Center in the Upper Yard Area.” Id. at 78270, 78313. Afterwards, more than 9,000 tons of clean backfill were imported to the site. In addition, a “surficial cap” was installed in the southern portion of the Upper Yard in order to “reduce the potential for mobilization of contaminants in the vadose zone to the groundwater.” Id. at 78287; see also Ex. 227 at 74424 (“Remedial activities ... include the excavation of a former gas holder, excavation of two tar vats, excavation of suspected source areas containing purifier waste and coal tar contaminated soils, and the surficial excavation of purifier waste and coal tar contaminated soils throughout much of the southern half of Parcel B.”). In 1999, a portion of the Lower Yard was capped to prevent further contamination from migrating. See Ex. 1937 at *272 90896. According to William Hoynack, Northeast Utilities Service Company’s (“ÑUSCO”) Manager for Environmental Audit and Remediation, Yankee Gas expended $2.5 million in the clean-up of the Norwalk MGP from 1994 through 1999. See Ex. 1939 at 52279 (Annual Report to the Department of Public Utility Control). 14 Based on this evidence, the Court finds as a fact that initiation of physical on-site construction of a remedial action occurred in September 1994 at the Norwalk MGP. The Court finds this for several reasons. First, the clean up efforts did not occur because of a release or threatened release of contaminated material into the environment. Instead, the action was precipitated by the construction of the new Yankee Gas Service Center in 1994 and the desire to remediate the Upper Yard before building new construction on top of it. See Ex. 2069 at 62 (Deposition of John Ripp); Ex. 527 at 90896 (1st Quarter 2008 Environmental Liability Report); Ex. 2247 (“Much of the Norwalk MGP property was remediated as part of the site development in 1995.”). The contamination itself had been present since at least 1960 when the Nor-walk MGP ceased its operations. The fact that there was no precipitating event, besides Yankee Gas’s own redevelopment plans, strongly supports the finding that the action was undertaken as part of a long-term strategy for cleaning up the site. Second, the way the project was planned and implemented suggests that it was intended to be a remedial action. The project was conducted according to an “MGP Site Remediation Strategy” that was endorsed by the Environmental Protection Agency (“EPA”) and implemented only after extensive investigations reaching back as far as 1989 into the extent and location of the contamination. See Ex. 231 at 78270, 78276. In other words, five years passed between the first site investigations and the initiation of the action. Rather than responding to an emergency situation, it is clear that Yankee Gas was implementing a voluntary clean-up plan approved by the EPA in anticipation of their redevelopment of the site. Third, the amount of money expended' — • $2.5 million — in relation to the estimated future remediation costs — between $3.5 million and $4.4 million — also supports the proposition that the action was the first phase of a remediation as opposed to a short-term clean-up arrangement. See Ex. 1939 at 52279 (Annual Report to the Department of Public Utility Control). As explained above, remedial actions generally take longer and are more expensive than removal actions. That was certainly the case here. Fourth, the Court finds it significant that Mr. Hoynack himself described the action at Norwalk as “remediation” and “remedial.” See Ex. 2047 (Memorandum from William J. Hoynack to Department of Public Utility Control) (“Significant remedial measures were performed at the Nor-walk and Willimantic sites from 1994 through 1999.”); Ex. 2110 (Deposition of William Hoynack) (“Q. So this remediation began in 1990s but it has to be continued now is that the idea? A. Remediation focused on the upper portion, correct.”). At trial, Mr. Hoynack admitted that he was using “remediation” as a term of art and was not using it in a colloquial way. See Tr. Transcript at 673. Plaintiffs expert, Dr. Shifrin, also stated that the Nor- *273 walk MGP had been “remediated in 1994 to prepare for the construction of a Yankee Gas service center.” Ex. 1; see also Ex. 2069 (Deposition of John Ripp) (“Like I said earlier, Norwalk had already gone through some remedial actions associated with the new development out there.”). Yankee Gas used similar language in its 2001 Annual Report to the Department of Utility Control (“Remedial activities were voluntarily completed on the southern half of Parcel B. As a result, approximately 6,700 cubic yards of contaminated soil were removed and treated off-site. Approximately $2,545,000 were spent in these remediation efforts.”); see also Ex. 231 (Site Remediation Documentation Report for the Norwalk MGP). Clearly, the people involved in the project considered it a remedial action. Although the repeated use of the word “remedial” by Plaintiffs and their experts is not dispositive, it does serve to confirm the Court’s view that what occurred at the Norwalk site was a remediation. Plaintiffs argue that all measures that occurred at the site were “preliminary” but it is clear that physical on-site construction had begun as early as September 1994. The Site Remediation Documentation Report lists excavators, curbing machines, front end loaders, cranes, bulldozers, and vibratory rollers as present and in use on the construction site. See Ex. 231. This is precisely the type of equipment that the Second Circuit found constituted physical on-site construction in Schaefer. See 457 F.3d at 203 (“While the mere purchase of a crane does not constitute the initiation of physical on-site construction, Schaefer’s subsequent use of the crane to spread cover (a mixture of topsoil, sand, and gravel) over his landfill satisfies each of the four statutory prerequisites.”). The measures taken were not mere sampling or investigation; instead nearly 9,000 tons of contaminated soil were removed. The cost of the action alone — $2.5 million— indicates that it was more than preliminary. Plaintiffs contend that the measures were preliminary because the Lower Yard and the Norwalk River still need to be remediated. However, the parties agree that each MGP constitutes a single “facility” under CERCLA. The Second Circuit interprets the term “facility” broadly to include land that is not owned by the plaintiff in a cost recovery action. See State of New York v. Shore Realty Corp., 759 F.2d 1032, 1043 n. 15 (2d Cir.1985) (“CERCLA defines the term ‘facility’ broadly to include any property at which hazardous substances have come to be located.”). Since there can only be one remedial action per facility, initiation of physical on-site construction of a remedial action in the Upper Yard started the clock for statute of limitations purposes. Plaintiffs argue that they should not be penalized for voluntarily initiating the clean up at these sites before they knew what the final clean up would entail and that strictly applying the statute of limitations would create perverse incentives for corporations facing CERCLA liability. While the Court is sympathetic to Plaintiffs’ position, the Court rejects this argument. Plaintiffs are sophisticated parties that should have known that beginning the remediation in 1994 would start the statute of limitations period. As explained above, they did not begin the clean up in response to an emergency situation that could not wait. Plaintiffs were in complete control of when the statute of limitations began to run. And there is no evidence that they were unaware that UGI might be a potentially responsible person under CERCLA. Furthermore, CERCLA provides a mechanism for precisely the situation Plaintiffs found themselves in. They could have brought a cost recovery action in 1994 and could have sought a declaratory judgment *274 for future costs associated with the cleanup of the Lower Yard and Norwalk River. See 42 U.S.C; § 9613(g)(2). They did not. Plaintiffs also point out that a remedial action must be consistent with a permanent remedy, see 42 U.S.C. § 9601(24), and that at least with respect to the surface cap that was installed in 1999, they will have to remove it at a later date. There is some contradictory evidence in the record on this point. For instance, according to the 1998 Annual Report to the Department of Public Utility Control, Yankee Gas decided not to install a “slurry wall and cap” because the Connecticut Department of Environmental Protection indicated that site closure would be unlikely in that case. Thus, Yankee Gas made the decision to install a surface cap in part because it wanted to move closer to site closure, see Ex. 2247 at 60348, indicating that it was at least contemplating a permanent remedy. However, even assuming that the surface cap that was installed in 1999 is temporary, the 1994 activities would constitute physical on-site construction of a remedial action. Mr. Hoynack admitted that the 1994 action was intended to be consistent with the permanent remedy and that there are no plans to return to redo it. See Tr. Transcript at 664, 678. Furthermore, the Court does not understand how the removal of 9,000 tons of contaminated soil could be inconsistent with the permanent remedy. Surely, Plaintiffs do not intend to return the contaminated soil to the site. And if that is the case, then any permanent remedy must be consistent with the soil removal. Because the Court finds as a fact that the remedial action at Norwalk was initiated in September 1994 and Plaintiffs filed suit more than six years after that date, their cost recovery action as to the Nor-walk MGP is time barred. Willimantic. According to Northeast Utilities’ First Quarter, 2008 Environmental Liability Report, an interceptor trench was installed at the Willimantic site in 1994 “to capture tars migrating to the river.” Ex. 2057 at 90902. In 1996, two subsurface holders were excavated. See id. In 1997, over 20,000 tons of contaminated soil were removed from the site. See id. These actions were taken, according to Mr. Hoynack, to prevent further migration or leaching of contamination from the site. See Tr. Transcript at 661. Yankee Gas expended over $6 million from 1992 to 2000 on these actions. See Ex. 1939 at 52279 (Annual Report to the Department of Public Utility Control). It is clear from this chronology — and the Court so finds — that physical on-site construction of the remedial action was initiated at the Willimantic MGP in 1994. The Court again says' this for several reasons. As with Norwalk, there was no release or threatened release of hazardous material that precipitated the action. Rather, Yankee Gas installed the trench in order to prevent migration of the hazardous materials that were already in the ground. See Tr. Transcript at 706 (Hoynack testimony) (“The trench was put in in the hope that that would intercept tar that was migrating from upgrading source to the river and be collected.”); Ex. 2069 at 63 (Deposition of John Ripp). “[P]revent[ing] or minimiz[ing] the release of hazardous substances so that they do not migrate to cause substantial danger to present or future public health or welfare or the environment” is one of the key objectives of a remedial action. 42 U.S.C. § 9601(24). In fact, the installation of trenches is specifically mentioned in the statute as an example of a remedial action. See id. Likewise, the excavation of the holders in 1996 and the removal of 20,000 cubic yards of contaminated soil in 1997 both constituted remedial activities. See id. (“The term [remedial action] includes ... cleanup of released hazardous substances and associ *275 ated contaminated materials, ... dredging or excavations, [and] repair or replacement of leaking containers.... ”). Moreover, as was the case with the Nor-walk MGP, Mr. Hoynack himself described the actions taken at Willimantie as “remedial” and “remediation.” See, e.g., Ex. 1938 (Memo from William Hoynack to Department of Public Utility Control) (naming “Past Remedial Steps” at Willimantie); Ex. 2047 (Memorandum from William J. Hoynack to Department of Public Utility Control) (“Significant remedial measures were performed at the Norwalk and Willimantic sites from 1994 through 1999.”); Ex. 2110 at 62 (Deposition of William Hoynack) (“Willimantie has had remediation work done under Yankee Gas. There was a focused effort to remove and clean up, I think, a relief holder, maybe two relief holders, some tar wells, and a bunch of contaminated soil that came out of there.”); id. at 83 (“Q: So those Willimantie costs, 170,000, 2.2 million, 3.3 million, those were remediation costs that had been incurred as you described earlier in the '90s? A: During the '90s, correct, by Yankee Gas, yes. That was older remediation, so on, so forth, for Willimantie.”). Yankee Gas also consistently characterized the action as remedial. The 1998 Annual Report to the Department of Public Utility Control states that site activities included the “ ‘[c]ompletion of a full scale remediation project of the Willimantie MGP property.” Ex. 2247 at 60344; see also id. (‘Yankee Gas completed the final excavation phases of remediation at the Willimantie site during the summer and fall of 1997. Yankee Gas achieved its goal of excavating two hot spots as defined by the Remediation Standard Regulations.”); id. at 60356 (“The 1996 remedial effort was a great success .... ”). Mr. Hoynaek’s and Yankee Gas’s characterization of the action as a remediation is consistent with the other evidence in the record. Plaintiffs again argue that the measures were “preliminary” or “interim” measures, but the Court rejects this argument. Yankee Gas expended over 95% of the cost of the cleanup between 1992 and 2000. According to Yankee Gas, the remainder of the cleanup will cost approximately $250,000 to $350,000 compared to the $6 million that has already been expended. See Ex. 1939 at 52279 (Annual Report to the Department of Public Utility Control). It would be odd, to say the least, if “preliminary” measures constituted 95% of the cost of the clean up. Similarly, the Court does not understand how Plaintiffs can argue that their actions were “interim” measures when Yankee Gas has estimated that only an additional $250,000 to $350,000 will be needed to complete the cleanup. If the permanent remedy cost so little, it would be nonsensical to spend $6 million on interim measures that were inconsistent with the permanent remedy. And again, Mr. Hoynack testified that the action was intended to be consistent with the permanent remedy: THE COURT: But I take it at least with respect to Willimantie, it may not have done all you wanted to do, but you — but whatever you did do previously is going to be consistent — is consistent with what you’re hope — you hope your final plan is going to be. In other words, it’s not that you did something that is just totally counter productive to what you hope the final plan will be. It’s a step in the process? MR. HOYNACK: It would be. The source material that’s there is more than likely going to have to come out. As with Willimantie. They focused on the source material. What they didn’t focus on is what remains in the soil. THE COURT: Right. But certainly nothing that they did in Willimantie is inconsistent with what you hope will be the final — part of the final— *276 MR. HOYNACK: That would probably be included in it, correct. See Tr. Transcript at 704. Likewise, Yankee Gas states in its 1998 Annual Report that “[t]he 1997 remedial effort was successful in that the site strategy and goal were achieved, putting the property closer to state closure and a No Further Action agreement.” Ex. 2247 at 60356. It is clear from this language that Yankee Gas also intended the action to be a part of the permanent remedy for the site. 15 Because the Court finds as a fact that physical on-site construction of the remedial action was initiated in 1994, more than six years before the Plaintiffs filed suit, the cost recovery action for the Willimantic MGP is time-barred. C. Plaintiffs make one further argument that the Court must consider with respect to the statute of limitations. They argue that it is “premature” for the Court to even consider the statute of limitations question because a final Remedy Action Plan (RAP) has not yet been approved for the sites and won’t be approved for many more years. They argue that it is impossible to know whether the actions are “consistent with permanent remedy,” as required under the statute, before the permanent remedy has even been chosen. Unfortunately for Plaintiffs, the Second Circuit has rejected the argument that a final RAP is required before the statute of limitations begins to run on remedial actions. In Schaefer, the Second Circuit held that “on-site construction” had begun when a crane was used “to dig, drag, and spread on-site soil and other material over the site as cover,” despite the fact that a final RAP had not yet been approved. 457 F.3d at 204. The Second Circuit’s interpretation is consistent with the statutory language, which states that the statute of limitations begins to run “after initiation of physical on-site construction,” 42 U.S.C. § 9613(g)(2)(b), not when the final RAP is approved. In fact, as the Second Circuit in Schaefer pointed out, if a final RAP were required, then the phrase “actions consistent with permanent remedy” in the definition of remedial action would be rendered superfluous. Plaintiffs attempt to distinguish Schaefer by arguing that “the activities at issue in Schaefer all occurred consistent with and in contemplation of the existing permanent remedy,” whereas here, according to Plaintiffs, “nobody can even say what the permanent remedy will be at Norwalk and Willimantic.” Pis.’ Posh-Trial Brief [doc. # 133] at 23. According to Plaintiffs, Schaefer merely rejected the application of a bright line rule requiring a final RAP before the statute of limitations is triggered. See Schaefer, 457 F.3d at 207. Plaintiffs are surely correct that if no one has any idea what the permanent remedy will look like, it would be impossible to determine whether the actions previously undertaken are consistent with the permanent remedy. After all, an action cannot be consistent with something that does not exist. However, Schaefer makes clear that courts may look at evidence other than a final RAP in order to determine what the permanent remedy is or will be. Here, it is simply not the case that “nobody can even say what the permanent remedy will be at Norwalk and Willimantic,” as Plaintiffs allege. *277 As detailed above, there is ample evidence that (1) Yankee Gas believed that the actions were consistent with the eventual permanent remedy; (2) the actions were developed in conjunction with the EPA and the State of Connecticut and pursuant to the remediation regulations then in effect; (3) in some cases, any final RAP must be consistent with the actions because they are irreversible (e.g., the removal of contaminated soil); (4) in the case of Willimantic, 95% percent of the expenses have already been incurred and there is very little work left to be done; and (5) in some cases, the measures have been in place for more than a decade and there are no plans to undo or remove them at this time. Moreover, there is no evidence in the record to suggest that the action undertaken were inconsistent with the permanent remedy. For Plaintiffs to argue, in the face of all of this evidence, that no one has any idea what the permanent remedy will be is essentially saying that a permanent remedy does not exist until the final RAP is approved. This is precisely the bright line test that the Second Circuit rejected in Schaefer. Therefore, the Court rejects Plaintiffs’ argument and concludes that the statute of limitations has run on the cost recovery actions at the Norwalk and Willimantic MGPs. V. In sum, UGI’s Motion for Judgment [doc. # 127] is GRANTED and judgment shall issue in favor of UGI on the cost recovery actions with respect to the following facilities: Norwalk, Bristol, Meriden Cooper Street, Middletown, Putnam, Rockville, Waterbury South, Winsted Gay Street, and Willimantic. If necessary, the Court will issue a scheduling order to proceed to the second phase of the trial with respect to Waterbury North. The parties are directed to file a joint status report with the Court by June 19, 2009. IT IS SO ORDERED. Footnotes [1] Although Yankee Gas is the current owner of many of the MGPs in question, the Court will focus — as did the parties — on CL&P. This is so because Yankee Gas acquired the MGPs after the time period in which UGI owned a majority stake in CL&P and therefore the corporate history of Yankee Gas is not relevant to Plaintiffs' CERCLA claims. The only fact about Yankee Gas that is relevant to this case is that it acquired some of the MGPs in question at some point between 1941 and the present and, therefore, is a potentially responsible person (“PRP”) under CERCLA. [2] Certain documents were no longer available, most notably, the minutes of the Works Committee of UGI from 1888-1904. Plaintiffs did not seek any adverse inferences against UGI because of the absence of documentation. [3] For the sake of clarity, TUGIC and UGI will be referred to collectively as UGI. UGI concedes that it is the successor corporation to TUGIC, although it disputes that TUGIC is the successor corporation to UGIC. Because the question of successor liability is relevant only to the Waterbury North MGP and UGI has conceded that it operated that facility, the Court reserves decision on the question of successor liability until the second phase of trial. [4] The Court will detail the ownership structure of each MGP in later sections. [5] Bestfoods involved an effort by the United States to recover under CERCLA the cost of remediating a facility near Muskegon, Michigan. In 1957, Ott Chemical Company had begun to manufacture chemicals at the facility. In 1965, the Bestfoods corporation (then known as CPC International, Inc.), created a subsidiary corporation that purchased Ott Chemical. The subsidiary corporation continued to manufacture chemicals until 1972, when Bestfoods sold it to an unrelated company. The district court held that Bestfoods was liable under CERCLA because it had "operated” the facility through its subsidiary. See CPC Int’l, Inc. v. Aerojet-Gen. Corp., 777 F.Supp. 549, 572 (W.D.Mich.1991). The district court reasoned that Bestfoods had operated the facility through its subsidiary because it had selected the subsidiary’s board and included its executives in the ranks of Bestfoods’s officers. Id. at 573. The Sixth Circuit reversed en banc. See United States v. Cordova Chem. Co., 113 F.3d 572 (6th Cir. 1997). It held, seven judges to six, that a parent corporation could only be liable for the actions of its subsidiary if the state common law rules for piercing the corporate veil were met or if the parent corporation operated the facility directly. See id. at 580. The Supreme Court vacated the circuit’s en banc opinion and remanded for further proceedings consistent with its opinion. [6] As noted later in this decision^ there is another problem with the claim regarding Nor-walk and that is that it is barred by the relevant statute of limitations. [7] In 1901, the Connecticut Lighting and Power Company changed its name to the Connecticut Railway and Lighting Company. To avoid confusion, the Court will refer to both entities as CR & L. [8] During the relevant time period, there were at times one or two holding companies be *248 tween UGI and CL&P. For ease of discussion and because there is no operator liability in any event, the Court will discuss the holding companies and CL&P as if they were one and the same. However, the Court makes no finding that the corporate veil of any of the holding companies should be pierced. [9] CL&P’s corporate histories (Ex. 1007, 1011, 1012) — which are very much contrary to the position CL&P is taking in this lawsuit — are not binding or judicial admissions, but they are evidentiary admissions that suggest to the Court that CL&P's current legal position is very much at odds with its long-term, and non-litigative, view of its relationship with UGI. See, e.g., Paice LLC v. Toyota Motor Corp., 504 F.3d 1293, 1312 (Fed.Cir.2007); Penguin Books USA, Inc. v. New Christian Church of Full Endeavor, Ltd., 262 F.Supp.2d 251, 259 (S.D.N.Y.2003); Guadagno v. Wallack Ader Levithan Assoc., 950 F.Supp. 1258, 1261 (S.D.N.Y.1997). [10] As UGI endeavored to show during the trial, Mr. Jebb wrote to many companies inquiring into their holder inspection practices as research for a presentation he was going to give at a conference. See Tr. Transcript at 348-55. [11] At his deposition, Dr. Shifrin testified that these holders were removed in 1929, which is before UGI's involvement with this facility. [12] Plaintiffs’ experts also cite a number of examples of UGI control between 1900-1906. The Court has already discussed these examples in its previous section on Norwalk. [13] UGI also raised a statute of limitations defense with respect to the Meriden Cooper Street and Winsted Gay Street facilities. The Court has concluded that it is unnecessary to decide whether the statute of limitations has run with respect to the removal operations conducted at these facilities. [14] ÑUSCO is an affiliate of Yankee Gas and CL&P. Northeast Utilities (“NU”) is the parent company of ÑUSCO, Yankee Gas and CL&P. Yankee Gas and CL&P are operating companies of NU. In 1989, NU divested from its gas businesses including Yankee Gas and in 2000, Yankee Gas was reacquired by NU. [15] Plaintiffs make the same argument that the need to remediate the Willimantic River somehow transforms what occurred into “preliminary” measures, but the Court rejects this argument for the same reasons it rejected the argument with respect to the Norwalk River.