This review of a summary judgment concerns the determinative factors for whether a corporation is required by its bylaw (indemnity, under certain conditions, for officers and directors sued “by reason of’ their corporate status) to indemnify an officer/director, Ralph L. Lowe, for his individual liability incurred under the Comprehensive Environmental Response, Compensation, and Liability Act, 42 U.S.C. § 9601 et seq. (CERCLA). Apparently because CERCLA permits personal liability to be imposed against an officer or director, the district court held that Lowe was not entitled to indemnity. We REVERSE.
*1007I.
This case arises out of the disposal of wastes at the Brio Superfund Site in Texas; the following facts are not in dispute. From the late 1950s until 1969, the site was owned by Hard-Lowe Company, and its successor Lowe Chemical Company (Lowe I). Lowe, an investor in Hard-Lowe, became the sole stockholder of Lowe I; and in 1969, he sold Lowe I to Chemical Pollution Control, Inc. (CPC). The transaction was financed by Lowe, who took a security interest in the Brio property. When CPC went bankrupt in 1972, Lowe foreclosed on the site.1
In anticipation of the foreclosure, Lowe incorporated Lowe Chemical Company (Lowe II). According to Lowe, he exchanged the Brio Site for stock in Lowe II in May 1972, but the attorneys handling the transaction “neglected to file and record a deed reflecting the conveyance of the property, and title to the property remained in [Lowe’s] name.” Due to this claimed oversight, Lowe was shown as the record owner of the site until 1977. Lowe claims that JOC discovered the error, and a “corrective deed” was recorded; it “recited that it was given by [Lowe] and accepted by JOC Oil Aromatics to ‘evidence the sale and intended conveyance made on May 2, 1972, and it shall be effective as of and retroactive to, such date.’ ”
According to Lowe, the purpose of the corrective deed related to the 1975 purchase of all of Lowe II’s stock by JOC at a price that assumed Lowe II’s ownership of the Brio Site. JOC defaulted on amounts owed Lowe, and conveyed the Brio Site back to him in 1978 in lieu of foreclosure. Later that year, Lowe conveyed the property to another entity.
After the Brio Site was declared a super-fund site, the government spent $1.31 million taking remedial action at it, and then filed an action against both JOC and Lowe to recover those costs. In turn, Lowe cross-claimed against JOC for indemnification under the corporate bylaw in issue. Without admitting liability, Lowe settled with the government for $400,000.2 JOC settled later for $20,000.
On Lowe’s cross-claim for indemnification, JOC and Lowe both moved for summary judgment. A magistrate judge recommended granting JOC’s motion and denying Lowe’s. After overruling Lowe’s objections to that recommendation, the district court entered judgment for JOC on Lowe’s cross-claim.
II.
Pursuant to Lowe II’s bylaws, each of its officers and directors was indemnified for liability and expenses incurred “in connection with any claim made against him, or any action ... to which he may be a party by reason of ... being” an officer or director.3 JOC and Lowe agree that this bylaw governs the indemnity claim; both assert here that they are entitled to summary judgment.4
*1008The district court held that Lowe was not sued “by reason of’ his corporate status. The magistrate judge found “it to be of significance” that the United States brought this action against Lowe personally, and held a hearing to ascertain whether the United States sued Lowe “in his capacity as an officer and director”. After examining the United States’ summary judgment submissions in the underlying litigation, and having discussions with a representative of the United States at the hearing, the magistrate judge concluded that Lowe was sued “because of actions he may have taken in his personal capacity and not because of actions he may have taken in his capacity as a director or officer.” This conclusion was driven by Lowe’s personal involvement at the Brio Site and the fact that CERCLA permitted personal liability for such actions.
A.
Based upon the complaint in the underlying litigation, and, alternatively, upon CERC-LA, JOC presents several bases for upholding its summary judgment.
1.
JOC seizes on the wording of the United States’ complaint, contending that the indemnity clause requires that Lowe be made a party to an action in his capacity as an officer or director. Thus, JOC urges that we focus on that complaint, and avoid asking “why or for whom Lowe was involved with hazardous substances at the Brio Site.” We disagree.
The parties agree that our interpretation of the indemnity clause is governed by Texas law; and under it, because the bylaw is not claimed to be ambiguous, its construction is a question of law. University Savings Ass’n. v. BuRNap, 786 S.W.2d 423, 425-26 (Tex.Ct.App.1990). Bumap addressed an issue very similar to that here: whether a corporate bylaw obligated University Savings to indemnify Burnap. Id. at 425. Burnap was an officer and director of University Savings’ predecessor, and he was sued by its former shareholders for “tipping inside information”. Id. at 424. Burnap requested indemnification for his legal fees incurred in defending the suit, pursuant to a bylaw that indemnified officers and directors for expenses incurred defending a suit “to which such person is made a party solely by reason of his being or having been a director [or] officer ”. Id. (emphasis added). The Texas Court of Appeals held that he was entitled to reimbursement. Id. at 426-27.
Interestingly, both parties rely on Bumap. Lowe maintains that it stands for the proposition that the pleadings are not the sole determinant for whether one is sued by reason of having been an officer or director; rather, the facts giving rise to the complaint must be considered. See id. at 426 (interpreting factual allegations made in complaint against Burnap and assessing trial testimony, among other things, in ascertaining whether indemnity clause applied). On the other hand, JOC contends that Bumap relied solely on what was apparent on the face of the complaint, and that the citation to trial testimony was “unnecessary” to the decision. See id. In fact, the decision did note that its holding “is consistent with the law concerning an insurance company’s duty to defend under an insurance contract. Under those provisions, the duty to defend is determined solely from the face of the pleadings in light of the policy provisions.” Id. (citation omitted).
We do not read Bumap to hold that indemnity is triggered solely by the underlying pleadings reciting that the individual is being sued as an officer or director. Obviously, this would give considerable power to artful pleading in the underlying complaint as to the application of the indemnity provision. Needless to say, this could, among other things, lead to considerable mischief. The examination of trial testimony and factual interpretation of the complaint’s allegations in Bumap do not appear to have been sur-plusage.
In any event, we need not rely solely on Bumap; we are guided by a recent deci*1009sion by the same court that decided Burnap, Grove v. Daniel Valve Co., 874 S.W.2d 150 (Tex.Ct.App.1994), application for writ of error filed (May 11, 1994).5 The court interpreted a Delaware statute that permitted indemnification when an individual is sued “by reason of the fact that he is or was a” director or officer, id. at 153 (quoting Del. Code Ann. tit. 8, § 145(a) (1991); emphasis deleted). As here, relying on Burnap, the appellee urged that the language of the complaint in the underlying litigation determined indemnification vel non. Id. at 152. The court found that reliance misplaced, ruling that “in Burnap we did not hold so sacred the alignment and naming of parties as to preclude any other analysis”, id. at 156:
Appellee’s argument proposes a static view of the litigation process, locking in indemnification rights at one stage of the process, namely at the preliminary stage of the proceedings. The pleading must be a primary part of any follow-up indemnification suit because it informs a court not involved in the original lawsuit of the substance of the suit, but it cannot be the end-all.
Id. at 158. Accordingly, we agree with the district court that the absence of a recitation in the United States’ complaint that Lowe was being sued because he was an officer and director of Lowe II is not fatal to his indemnification claim. But, as discussed infra, we part company with the district court as a result of the determinative effect it apparently gave to CERCLA providing for personal liability for actions taken by Lowe. While CERCLA does permit such liability, this is not the end of the inquiry; a more detailed factual inquiry was required as to Lowe’s status during the relevant period.
2.
JOG asserts that Lowe was personally involved in the operation at the Brio Site, and thus liable under CERCLA.6 See 42 U.S.C. § 9607(a)(2) (’'mposing liability on “any person who at the time of disposal of any hazardous substance owned or operated any facility at which such hazardous substances were disposed of’); Riverside Mkt. Dev. Corp. v. International Bldg. Prods., Inc., 931 F.2d 327, 330 (5th Cir.) (per curiam) (“CERCLA prevents individuals from hiding behind the corporate shield when, as ‘operators,’ they themselves actually participate in the wrongful conduct prohibited by the Act.”), cert. denied, — U.S. -, 112 S.Ct. 636, 116 L.Ed.2d 654 (1991); United States v. Northeastern Pharmaceutical & Chem. Co., 810 F.2d 726, 743-44 (8th Cir.1986) (government can impose personal liability on an officer or employee of a corporation as a “person” within the meaning of CERCLA), cert. denied, 484 U.S. 848, 108 S.Ct. 146, 98 L.Ed.2d 102 (1987). JOC contends that Lowe’s liability under CERCLA precludes his liability arising “by reason of’ being an officer and director.
We disagree. Although Lowe’s liability “was not derivative but personal”, see Northeastern Pharmaceutical, 810 F.2d at 744, this does not preclude, per se, his being indemnified. A similar argument was attempted in one of the few reported federal decisions addressing whether an indemnity clause covered personal CERCLA liability, Bowen Engineering v. Estate of Reeve, 799 F.Supp. 467 (D.N.J.1992). There, the estate of a former officer sought indemnification from the corporation, which had joined other corporations in .suing the estate for contribution for CERCLA damages. In that case, the corporate plaintiffs “contend[ed] that they [were] suing [the former officer], through his estate, personally for the violation of CERCLA.” Id. at 485 (emphasis in original). Nevertheless, the court concluded that although the officer (and therefore his estate) was Hable personally under CERCLA as an “operator” of the site, id. at 474, he was entitled to indemnification under a clause nearly identical to that in issue here. See id. at 483-89.7
*1010We agree with Bowen. Lowe’s CERCLA personal liability does not, without more, void the indemnity clause. In other words, although Lowe is personally liable, he may be entitled to indemnity. Cf. Grove, 874 S.W.2d at 154 (“It is true that ... Grove could be individually liable for any negligent de-sign_ The indemnification statute, however, expressly provides that if an employee’s job exposes him to liability, then the corporation must indemnify him if the employee is successful on the merits....”) Indeed, if Lowe II’s officers or directors were not subject to personal liability for acts or omissions arising out of their corporate positions, there would be little reason for the indemnity clause. Cf. Riverside, 981 F.2d at 380 (“ ‘Corporate officers are liable for their torts, although committed when acting officially.’ ”) (quoting 3A S. Flanagan & C. Keat-ing, Fletcher Cyclopedia of the Law of Private Corporations § 1135 (1986)). As stated, Lowe’s CERCLA personal liability does not preclude per se indemnification.8
3.
Alternatively, JOC contends that the bylaw’s prohibition of indemnification where such “would be against public policy” prohibits it in this instance, asserting that “[a]llowing Lowe to shirk his responsibility for the contamination he caused violates CERCLA’s policy.”
But, assuming arguendo that this issue was raised properly in district court, JOC does not contend that CERCLA prohibits indemnification; to the contrary, it notes “that 42 U.S.C. § 9607(e)(1) freely allows parties to agree to indemnify one another for liability under CERCLA.” That being so, we cannot understand why Lowe’s personal liability under CERCLA would be contrary to CERCLA’s allowing indemnity. Obviously, indemnification does not reheve Lowe of his Lability vis-a-vis the government; it only provides a possible source of payment for damages resulting from that liability.9
B.
As noted, Lowe does more than merely challenge the adverse summary judgment; he asserts that he is entitled to judgment.10 But, as noted earlier, more factual inquiry into the relationship between Lowe’s CERCLA liability and his status as an officer and director is required.
The focus of that factual inquiry should be on the “connection between the complaint and [Lowe’s] corporate status”. See Grove, 874 S.W.2d at 156. In assessing this connection, we recognize the breadth with which many courts have interpreted language such *1011as “by reason of’. For example, the Seventh Circuit recently interpreted the Delaware indemnification statute, and concluded that the “ “by reason of the fact that’ phrase is broad enough to encompass suits against a director in his official capacity as well as suits against a director that arise more tangentially from his role, position or status as a director.” Hefferman v. Pacific Dunlop GNB Corp., 965 F.2d 369, 375 (7th Cir.1992).11
Keeping the breadth of the bylaw’s “by reason of’ language in mind, Lowe’s liability may be connected to his status as an officer and director; much of his liability may have arisen “by reason of’ that status. Even JOC offers support for this position. For example, it states that “Lowe personally oversaw the day-to-day site activities, was actively involved in Lowe Chemical Company’s environmental compliance problems, supervised the initial pit closure activities, and was involved in and had the authority to control the disposal of waste generated by Lowe Chemical Company.” (Footnote omitted.) Similarly, JOC recognizes that “Lowe personally participated in the disposal of styrene tar at the Brio Site and maintained daily, on-site involvement in decisionmaking regarding hazardous substances at Lowe Chemical Company — a company whose fundamental operations involved the use of hazardous substances it placed in earthen pits.” (Footnote omitted.)
Obviously, a number of other issues require exploration by the district court. First, as discussed, Lowe was sued as both an “operator” and an “owner” of the Brio Site. Lowe suggests that, to the extent his liability arose as an “owner” rather than an “operator”, “this theory [of liability] does not entitle [him] to indemnification.” Therefore, there may be a need for apportionment of Lowe’s liability between that occasioned as an “owner” and that as an “operator”.12
Moreover, Lowe contends that he was not liable at all as an “owner” of the Brio Site. If this is true, apportioning the liability may be pretermitted (assuming apportionment is appropriate). Lowe advances two reasons for escaping “owner” liability. First, he asserts that he is exempt from CERCLA liability as a foreclosing lienholder. See 40 C.F.R. § 300.1100 (1993). Second, he contends that JOC’s acceptance of the “corrective” deed for the Brio Site estops it from asserting that Lowe owned the Brio Site during the years preceding that deed. These issues are appropriately first addressed by the district court.
Last, assuming apportionment is required, the question of attorneys’ fees arises. Lowe notes that the indemnity clause covers “expenses ... incurred ... in connection” with claims subject to the clause. He contends that expenses include attorneys’ fees, and that they should not be subject to apportionment. More specifically, he analogizes the expense provision to an insurer’s duty to defend. Under Texas law, “the insurer is obligated to defend, as long as the complaint alleges at least one cause of action within the coverage of the policy.” Rhodes v. Chicago Ins. Co., 719 F.2d 116, 119 (5th Cir.1983) (citations omitted); see also Burnap, 786 S.W.2d at 426 (inviting comparison between indemnity clause and insurer’s duty to defend). On the other hand, Grove directed that, on remand, “the parties ... apportion in some reasonable manner the monies spent by [Grove] defending his status as general partner [for which indemnification was not allowed] and as corporate personnel [for which it was].” Grove, 874 S.W.2d at 158. We need not clarify Texas law on' this issue; until the district court determines whether apportionment of the liability is necessary, and, if so, how it should be made, there is no issue before us regarding apportionment of attorneys’ fees. In short, this is another possible issue for the district court.
III.
For the foregoing reasons, the judgment is REVERSED, and the case REMANDED for *1012farther proceedings consistent with this opinion.
REVERSED and REMANDED.