HARRISON L. WINTER, Chief Judge:
Plaintiffs Chesapeake Bay Foundation (“CBF”) and National Resources Defense Council (“NRDC”) filed a “citizen suit” under section 505 of the Clean Water Act, 33 U.S.C. § 1365, against defendant Gwaltney of Smithfield, Inc. (“Gwaltney”), alleging violations of the pollutant effluent limits contained in Gwaltney’s National Pollutant Discharge Elimination System (“NPDES”) permit, issued pursuant to section 402 of the Act, 33 U.S.C. § 1342. The district court granted plaintiffs’ motion for partial summary judgment on the issue of liability and held a hearing to determine the amount of the civil penalty to be assessed against Gwaltney pursuant to section 309(d) of the Act, 33 U.S.C. § 1319(d). Several months later, Gwaltney filed a motion to dismiss for lack of subject matter jurisdiction, alleging that it had ceased violating its permit prior to plaintiffs filing suit. Gwaltney argued that citizen suits for purely past violations are not permitted under the Act, and that for jurisdiction to lie, the Act requires a defendant to be violating the Act at the time suit is filed.1
The district court denied the motion to dismiss, holding that citizen suit jurisdiction could lie in the absence of an ongoing violation. It also ruled that Gwaltney was liable for a maximum civil penalty2 of $6,660,000, but, in view of a variety of factors, adjusted the penalty downward to $1,285,322.3 Gwaltney appeals, and we affirm.
I.
Plaintiffs are nonprofit corporations dedicated to protecting natural resources. CBF is a regional environmental group with over 19,000 members residing in the Chesapeake Bay area, and NRDC a nationwide environmental group with over 800 of its members residing in Virginia. Gwalt-ney is a subsidiary of Smithfield Foods, Inc., and is engaged in the business of processing and packing pork products. Smithfield Foods had acquired the Gwalt-ney plant, which is situated on, and discharges wastewater into, the Pagan River near Smithfield, Virginia, from the ITT-Continental Baking Co. Gwaltney assumed responsibility for wastewater discharge under the NPDES permit as of October 27, 1981; although numerous violations of the permit occurred before that date, only those violations occurring afterward are the subject of this lawsuit.
Jurisdiction over this citizen suit was grounded on Section 505(a) of the Act, 33 U.S.C. § 1365(a), which permits any citizen to commence a civil action against any person alleged to be in violation of an effluent standard or limitation under the Act, in order to enforce the limitation and/or to assess civil penalties for its violation.4 The [307]*307suit was filed more than sixty days after plaintiffs, in compliance with section 505(b)(1)(A) of the Act, 33 U.S.C. § 1365(b)(1)(A), had given Gwaltney, the Environmental Protection Agency (“EPA”) and the Virginia State Water Control Board notice of Gwaltney’s violations and of the plaintiffs’ intent to file suit.
The violations involved in this case (the responsibility for which is not at issue on appeal) are repeated discharges in excess of Gwaltney’s NPDES permit limits for a number of pollutants.5 That permit fixed both daily and “monthly average” limits for certain named pollutants. The most significant pollutants were chlorine (Cl2), for which the last violation occurred in October 1982, and total Kjeldahl nitrogen (TKN), for which the last violation occurred on May 15, 1984. The latter was the last recorded violation for any pollutant and occurred approximately one month before this suit was filed. These violations were reported by Gwaltney itself, as required by law, see 33 U.S.C. § 1318(a)(3)(A); 40 C.F.R. § 122.41(f)(4), in its discharge monitoring reports (DMR’s).
Many of the excessive discharges were violations of daily limits for the various pollutants. Others were violations of “monthly average” limits for which measurements are made on the basis of a limited number of samples taken during the month, according to the terms of the NPDES permit.6 For purposes of assessing civil penalties under 33 U.S.C. § 1319(d), which provides for a maximum penalty of $10,000 per day of a permit violation, the district court treated each monthly average violation as if it were a series of daily violations, one for each day of the month in question. Thus the court found monthly average violations to have occurred in twenty-two different months, for a total of 653 days of violation.7 In addition, the district court assessed penalties for thirteen violations of daily pollutant limits, giving a grand total of 666 days of violation,8 and a potential maximum pen[308]*308alty of $6,660,000. It then adjusted downward the penalties actually to be imposed and fixed an aggregate penalty of $1,285,-322. In making the adjustment, the district court, with the agreement of the parties, used as a guideline the EPA’s Civil Penalty Policy, 41 Env’t Rep. (BNA) 2991 (Feb. 16, 1984). In accordance with EPA’s policy, the court considered such factors as economic benefit derived by Gwaltney from noncompliance, a “gravity” component based on the seriousness of the particular violation, and unwarranted delays in compliance.
II.
A. Subject Matter Jurisdiction
Because it is undisputed that Gwaltney had ceased violating its permit before June 15, 1984, when suit was instituted, Gwalt-ney contends that the district court lacked jurisdiction over the subject matter of the suit. It argues that the plain language of the Act requires this result, because a defendant, under section 505(a), must be “in violation” of the Act for jurisdiction to attach.9 Gwaltney interprets this statutory language as requiring a polluter to be violating the Act at the time suit is filed against it. Gwaltney further contends that the structure of the statute and its legislative history limit the role of citizen suits to seeking abatement of ongoing violations and obtaining civil penalties for past violations only when jurisdiction is established by the existence of an ongoing one. We disagree with each of these contentions and hold that the Clean Water Act authorizes citizens, as “private attorneys general,” see Middlesex County Sewerage Auth. v. National Sea Clammers Ass’n, 453 U.S. 1, 14 n. 23, 17, 101 S.Ct. 2615, 2623 n. 23, 2624 n. 23, 69 L.Ed.2d 435 (1981), to seek civil penalties for past violations of the Act as well as abatement of ongoing violations.
1. Statutory Language
Gwaltney argues that the ordinary meaning of section 505 requires that a de[309]*309fendant be engaged in a violation at the time the complaint against it is filed for jurisdiction to lie. See Hamker v. Diamond Shamrock Chemical Co., 756 F.2d 392, 395 (5 Cir.1985). The Hamker court refused to read “to be in violation” as meaning “to have violated,” stating that such an interpretation “obviously strains the grammar of the statute and diverges from its ordinary meaning.” Id.
We respectfully disagree. We do not read the language as encompassing only those situations in which a defendant is currently violating the Act. Bather, we agree with the district court that the language is ambiguous, in that it can be read to comprehend unlawful conduct that occurred only prior to the filing of a lawsuit as well as unlawful conduct that continues into the present.10 See also Student Public Interest Research Group v. Monsanto Co., 600 F.Supp. 1474, 1476 (D.N.J.1985) (“A plausible construction of the language is that one is ‘in violation’ and continues to be ‘in violation’ by having ‘violated.’ In other words, the taint of a past violation is continuing.”). We therefore cannot, as Gwaltney urges, rely on the “plain meaning” of this language in deciding this case. Instead, we look beyond the language itself to the structure of the statute and its legislative history to discern congressional intent as to the scope of citizen suit jurisdiction.
2. Statutory Structure
In light of the language of numerous other sections of the Act, Gwaltney’s reliance on the words “in violation” to curtail the scope of citizen suits undermines its position. Virtually all of the enforcement provisions of the Act, for citizen and government enforcement alike, employ similar, present-tense phrasing. See, e.g., 33 U.S.C. § 1319(a)(1) (EPA Administrator authorized to take enforcement action, including court action, whenever he “finds that any person is in violation of any condition or limitation_”); id. § 1319(a)(3) (Administrator may issue a compliance order or file a civil suit whenever he “finds that any person is in violation” of certain provisions of the Act or of permit conditions or limitations); id. § 309(c)(1) (criminal penalties may be imposed on “[a]ny person who willfully or negligently violates” certain provisions of the Act or permit conditions or limitations). It can hardly be questioned that the EPA has authority to bring suit for civil penalties for purely past violations. Cf. United States v. Earth Sciences, Inc., 599 F.2d 368, 375-76 (10 Cir.1979) (EPA may, under 33 U.S.C. § 1319(a)(3), either issue a compliance order to prevent future pollution or seek civil penalties for past violations, or do both); United States v. Detrex Chemical Industries, Inc., 393 F.Supp. 735, 738 (N.D.Ohio 1975) (same); Student Public Interest Research Group v. AT & T Bell Laboratories, 617 F.Supp. 1190, 1197 (D.N.J.1985) (“[Tjhere is no doubt that the EPA may bring suit based on past violations.” Student Public Interest Research Group v. Monsanto Co., 600 F.Supp. 1474, 1476 (D.N.J.1985) (“It can hardly be argued that the Government is restricted to abatement actions [to the exclusion of suits for civil penalties]”). Thus “in violation,” in the context of government enforcement, has been construed to encompass past, completed violations as well as current, on-going transgressions. Any other reading of the Act would eliminate a significant deterrent to violations of the Act and severely undercut the Act’s ambitious purpose, “to restore and maintain the chemical, physical and biological integrity of the nation’s waters,” 33 U.S.C. § 1251(a).
Similarly, a significant deterrent would be lost if citizen suits seeking civil penalties for past violations were not per[310]*310mitted. For this reason, the scope of citizen enforcement powers must, in this important respect, be viewed as co-extensive with the enforcement powers of the EPA. A number of courts have so held, based largely on the similarity of language between the provisions authorizing government units and the citizen suit provision. See Connecticut Fund for the Environment v. The Job Plating Co., 623 F.Supp. 207, 213 (D.Conn.1985) (“[Rjemedies obtainable in citizen suits should be co-extensive with those available in suits initiated by the federal government.”); Student Public Interest Research Group v. Georgia-Pacific Corp., 615 F.Supp. 1419, 1425 (D.N.J.1985) (“In suits under Section 505 of the Clean Water Act, citizens have the same remedies available to the EPA”); AT & T Bell, 617 F.Supp. at 1199 (“[T]o hold that citizen suits alone are barred from seeking civil penalties for past violations would thwart the goal of uniformity in enforcement.”).
Gwaltney argues further that the Act is structured so that citizen suits are to play a “distinct and limited” role in the enforcement of the Act, one “supplementary” to the roles of state and federal agencies. While it cannot be denied that citizens’ enforcement authority is narrower than that of government — citizens, for example, obviously cannot seek criminal penalties or issue compliance orders — the Act does not deprive citizens of the right to sue for past violations. This conclusion is supported by the fact that the Act does impose certain express limitations on the bringing of citizen suits. Section 505(b), 33 U.S.C. § 1365(b), provides that no citizen suit may be commenced less than sixty days after the plaintiff has given notice of an alleged violation to the Administrator, the state in which the violation occurs, and the alleged violator. The section also prohibits filing of a citizen suit if either the Administrator or the state has commenced and is diligently prosecuting a civil or criminal action in federal or state court. In such cases, however, a citizen is permitted to intervene as a matter of right. These provisions have been narrowly construed by the courts, e.g., Friends of the Earth v. Consolidated Rail Corp., 768 F.2d 57, 63 (2 Cir.1985) (initiation of administrative enforcement action by state agency did not preclude filing'of citizen suit), and the Act contains no other express limits on institution of citizen suits. We read the statute as requiring only those limits on citizen suit jurisdiction that Congress expressly provided; we see no reason to impose by implication limits which Congress could have, but did not, create.
The fact that section 505(a) itself expressly authorizes the court in a citizen suit to “apply any appropriate civil penalties under [section 309(d), 33 U.S.C. 1319(d)]” lends further support to plaintiffs’ position. Courts have relied on this fact in concluding that the Act authorizes citizen suits seeking civil penalties for past violations. See, e.g., Georgia Pacific, 615 F.Supp. at 1425 (language of sections 505(a) and 309(d) “has been held to sustain civil penalties for past violations”) (citing United States v. Earth Sciences, Inc., 599 F.2d 368, 376 (10 Cir.1979); Monsanto, 600 F.Supp. at 1476 (section 505(a) “quite specifically refers to the court’s power to impose civil penalties and contains no limiting time frame”). Student Public Interest Research Group v. Anchor Thread Co., 22 Env’t Rep.Cas. (BNA), 1150, 1154 (D.N.J.1984); Job Plating, 623 F.Supp. at 213.11
[311]*311Moreover, an expansive reading of the jurisdictional grant contained in section 505(a) is consistent with the recognition by Congress and the courts of the importance of citizen suits as an enforcement tool. See S.Rep. No. 414, 92d Cong., 1st Sess. 79 (1971), U.S.Code Cong. & Admin.News 1972, pp. 3668, 3745, reprinted in 1 A Legislative History of the Water Pollution Control Act Amendments of1972, at 1497 [hereinafter cited as Legislative History] (“The Committee has established a provision in the bill that would provide citizen participation in the enforcement of control requirements and regulations under the Act modeled on the Clean Air Amendments of 1970.... One modification would allow the courts to impose civil penalties provided as a result of actions brought by citizens.”); Clean Water Act Amendments of 1985, Senate Environment and Public Works Comm., S.Rep.No. 50, 99th Cong., 1st Sess. 28 (1985). (“Citizen suits are a proven enforcement tool. They operate as Congress intended — both to spur and supplement government enforcement actions. They have deterred violators and resulted in significant compliance gains.”); AT & T Bell, 617 F.Supp. at 1199 (fact that Middlesex court viewed citizens as “private attorneys general” under the Act, “suggest[s] that Congress intended citizens to step into the shoes of government agencies that failed to act”),
Given this recognition of the importance of citizen suits seeking civil penalties, and the fact that Congress knew how to, and did, in some ways, expressly limit the application of these suits, there is no reason to suppose that Congress intended to impose an additional restriction — the requirement of an ongoing violation — not expressly provided for and not imposed on civil suits by the government.
3. Legislative History
Gwaltney also contends that the legislative history of the Act firmly supports its position. The fact is that the legislative history does not conclusively support either side’s position, but to the extent it speaks to the jurisdictional issue before us, we conclude that it favors the plaintiffs.
Gwaltney relies almost exclusively on passages from the Act’s legislative history that emphasize the use of citizen suits to abate ongoing violations. See, e.g., S.Rep. No. 414, 92d Cong., 1st Sess. 79-81, reprinted in 2 Legislative History at 1497-98. The cited passages, however, do not state that abatement of ongoing violations is the only relief available under section 505(a), and in fact refer to the availability of civil penalties. Id. at 1497.12
The sections of the legislative history that speak most directly to the issue before us are the oral and written statements of Senator Muskie,13 one of the principal authors and sponsors of the bill, commenting on the Conference Committee’s conclusions, which he presented on the Senate floor. Most notably, Mr. Muskie stated: “As in the original Senate bill, a citizen has the right under section 505 to bring an action for an appropriate remedy in the [312]*312case of any person who is alleged to be or to have been, in violation, whether the violation be a continuous one, or an occasional or sporadic one.” 118 Cong.Rec. 33700, (1972), reprinted in 1 Legislative History at 179 (emphasis added). The Senator also noted that “[c]itizen suits can be brought to enforce against both continuous and intermittent violations.” Id. at 163. Mr. Muskie’s comments indicate quite strongly that the Act was intended to permit citizens, as private attorneys general, to sue violators even in the absence of a violation that is ongoing at the time suit is filed.
4. The Hamker Decision
Gwaltney relies heavily on Hamker v. Diamond Shamrock Chemical Co., 756 F.2d 392 (5 Cir.1985). As suggested in Section IIA(l) above, we find the reliance to be misplaced. Hamker involved a citizen suit seeking civil penalties, an injunction, and damages by way of pendent state law claims, as a result of an oil spill, a one-time occurrence that took place many months before suit was filed. There was no effluent permit or compliance order involved. The Hamker court relied chiefly on the “ordinary meaning” of the statute and on Middlesex County Sewerage Auth. v. National Sea Clammers Ass’n, 453 U.S. 1, 101 S.Ct. 2615, 69 L.Ed.2d 435 (1981), in concluding that prospective relief is the only kind of relief authorized by the Act’s citizen suit provision, and that civil penalties are available “only as prospective relief.” 756 F.2d at 398-99. Because no permit violation was involved in Hamker, because only a single long-past and non-recurring discharge (for which not even a good faith allegation of a possible continuing violation could have been made), and because the Hamkers failed to allege a violation of any effluent standard or limitation under the Act, it could be argued that Hamker is distinguishable on its facte and on that basis inapplicable to the case before us. But even if not distinguishable, we decline to follow it to the extent that it holds that section 505(a) of the Clean Water Act does not permit citizen suite seeking civil penalties for past violations. As we have discussed above, we disagree that the “ordinary meaning” of the statute requires the adoption of the rule espoused in Hamker. The language “to be in violation of” is at least ambiguous, and our reading of the statute as a whole and its legislative history lead us to interpret that language as permitting citizen suite for violations occurring solely in the past.
In addition, we find the reliance by the Hamker court, and by the defendant in the case before us, on the Supreme Court’s opinion in Middlesex to be misplaced. The issue in Middlesex was whether the Clean Water Act, with its “elaborate enforcement provisions,” 453 U.S. at 13, 101 S.Ct. at 2623, authorized an implied private cause of action for damages as well. The Court held that it did not. Any statements that might be read as limiting the scope of citizen suits to prospective relief were mere dicta.14 The Court, in fact, acknowledged the availability of civil penalties under the Act. Id. at 14 n. 25, 101 S.Ct. at 2623 n. 25.
A final comment is required concerning Gwaltney’s expressed concern that to allow suite such as the one at bar may result in [313]*313the inundation of federal courts with Clean Water Act citizen suits, contrary to the underlying policy of the Act. The Hamker court voiced a similar concern:
If section 1365 were interpreted as permitting citizen suits for civil penalties for past violations, all state damage claims which could be brought under pendent jurisdiction could be litigated in a federal forum, thus undermining congressional intent to limit the burden on the district courts.
756 F.2d at 396. The concern is a real one, but does not require us to place undue restrictions on the scope of citizen suits under the Act. The wary exercise of pendent jurisdiction by the district courts could limit the burden imposed by citizen suits on the federal court system. While the district courts have the power, by invoking pendent jurisdiction, to hear state law claims asserted in federal court, the authority to exercise that power is discretionary. See United Mine Workers v. Gibbs, 383 U.S. 715, 726, 86 S.Ct. 1130, 1139, 16 L.Ed.2d 218 (1966) (“It has consistently been recognized that pendent jurisdiction is a doctrine of discretion, not of plaintiff’s right.”) In the case of the Clean Water Act, under which it was the intent of Congress to preclude suits for damages, see Middlesex, 453 U.S. at 18, 101 S.Ct. at 2625, a district court would be acting within its discretionary authority in refusing to exercise jurisdiction over state law damages claims. This would be particularly true in cases such as Hamker, where it is apparent that a plaintiff’s claim is predominantly a state law damages claim, with the Clean Water Act serving merely as a vehicle for bringing that claim in federal court. Refusal to entertain such claims would eliminate the lure of obtaining damages in federal court and thus avoid the “floodgates” problem envisoned by the Hamker court.
In sum, we hold that citizen suits like the one at bar, seeking civil penalties for permit violations committed entirely in the past, are permitted under section 505(a). Gwaltney has not asserted, and indeed could not assert, given section 505(a)’s express provision for “appropriate civil penalties under section [309(a), 33 U.S.C. § 1319(d) ],” that civil penalties may never be assessed in a citizen suit. However, the narrow reading of section 505(a) that it urges, creating jurisdiction to impose such penalties only in suits involving violations that continue up to the time suit is filed, places an untenable limitation on the use of citizen suits and civil penalties as a tool of enforcement and deterrence.
B. Gwaltney’s Violations of Monthly Average Limitations
Gwaltney argues that the district court committed reversible error when, in computing a potential maximum civil penalty pursuant to 33 U.S.C. § 1319(d), it treated each violation of a monthly average limitation as equivalent to a daily violation for each day of that month. Gwaltney suggests instead that a monthly violation be treated as a single day of violation. We disagree.
There is very little authority addressing this issue. Neither party cites any administrative regulatory materials that direct us in assessing penalties under 33 U.S.C. § 1319(d) for exceeding a monthly average limitation, and we have found none. In the one reported case that has dealt with this matter, the court was not faced with precisely the same issue as is before us. In United States v. Amoco Oil Co., 580 F.Supp. 1042 (W.D.Mo.1984), the EPA sought civil penalties for numerous violations of monthly average limitations in Amoco’s permit. Amoco argued that, because such average limitations could not be related to a specific, given day, no civil penalty could be imposed under 33 U.S.C. § 1319(d). Because Amoco’s permit contained no daily maximum limitations, Amoco would have been subject to no penalty whatsoever if the court had accepted its argument. Faced with this alternative, the court held that Amoco was subject to a penalty for each day of each month in which a monthly average limitation was exceeded. In the case before us, the district court could have assessed penalties [314]*314for numerous daily violations without resorting to the more extreme approach followed by the Amoco court.
Although the Amoco court may have, as Gwaltney argues, gone beyond what was necessary to decide the issue before it, we nevertheless concur in that court’s reasoning and interpretation of 33 U.S.C. § 1319(d).
We think that the language of the statute supports this view. While the statute does not address directly the matter of monthly average limitations,15 it does speak in terms of penalties per day of violation, rather than penalties per violation. This language strongly suggests that where a violation is defined in terms of a time period longer than a day, the maximum penalty assessable for that violation should be defined in terms of the number of days in that time period. We agree with the district court, that “[i]t is difficult ... to imagine how the violation of a monthly limitation involves any number of days other than the number of days in that month.” 611 F.Supp. at 1552. We certainly must reject the view of the defendant in Amoco, that violation of a monthly limitation should be treated as no violation at all. It makes little more sense to treat a monthly violation as one day of violation, as Gwalt-ney would have us do. This approach, in essence, sets a maximum penalty per violation, rather than per day of violation, a result which we see as inconsistent with the language of § 1319(d).
Moreover, the approach employed by the Amoco court and the district court in the instant case is essential to providing a district court a framework within which it will have sufficient flexibility to assess penalties that suit the particular circumstances of each case. A hypothetical offered by the district court serves to illustrate this point. The court’s hypothetical polluter very nearly exceeds its daily maximum every day of the month. This polluter would far exceed its monthly average limitation,16 and would be liable for violating that limitation. According to Gwaltney’s view, however, a court would be required to treat such conduct as one single day of violation. This would prevent the court from imposing over $10,000 in penalties for a full month of substantial discharges. We, like the district court, do not find this result sensible. The illogic of the approach is further illustrated by considering a situation very similar to the district court’s hypothetical: The polluter just barely exceeds its daily maximum for each day of a month. This polluter would be subject to liability for thirty days of violation (a maximum penalty of $300,000), although its total discharge is only slightly more than the first polluter, who, according to Gwaltney’s view, would be subject to only one day’s penalty ($10,000 maximum). Gwaltney’s approach thus could act to deprive a district court of the flexibility to deal with similar cases similarly, imposing a drastic limitation on the court’s authority to assess “appropriate” civil penalties.
Gwaltney has offered its own hypothetical in an effort to expose the unfairness of the approach employed by the district court. Gwaltney imagines a polluter which discharges, for twenty-nine days of a month, 98% of the amount it would have to average each day to exceed its monthly limit. On one day, however, the polluter discharges 180% of that amount (the latter amount still being less than the daily maximum limitation). This polluter has violated its monthly average limitation, and, according to the district court’s view, would be subject to penalties for thirty days of violation (a $300,000 maximum penalty), even [315]*315though it never once exceeded its daily maximum.17
The seeming unfairness of the situation in Gwaltney’s hypothetical dissipates when one recalls that § 1319(d) serves only to set a maximum penalty. Within this limit, the district court in its discretion may set an “appropriate” penalty. A district court, applying the approach to penalty assessment that we adopt, could (and indeed should, all else being equal) impose a substantially smaller penalty on Gwaltney’s hypothetical polluter than on either of the polluters in the first two hypotheticals discussed.
Again, what is important is to adopt an approach that will give district courts the continuity of possibilities necessary for them to assess appropriate sanctions in every case. The district court’s approach achieves this end.
C. Amount of Penalty
Gwaltney contends that, in a number of ways, the district court abused its discretion and exceeded its statutory authority in establishing the amount of the penalty in this case.
First, Gwaltney argues that, in assessing a part of the penalty, the court impermissibly shifted the burden of proof to Gwaltney. The district court found that Gwaltney was responsible for thirteen violations of daily limitations occurring during months in which the monthly average limitation was not violated. Gwaltney’s DMR’s did not show on precisely which days the violations occurred; they could conceivably have occurred on as few as seven days. See 611 F.Supp. at 1556 & n. 15. The court presumed that they occurred on thirteen separate days, leaving it to Gwaltney to demonstrate through records solely within its control, that there was some “overlap”18 of violations, thus reducing the total number of days of violation. Because Gwaltney failed to come forward with evidence showing overlapping violations, the district court relied on the presumption that thirteen separate daily violations had occurred.
We find no infirmity in the district court’s approach. Plaintiffs, by showing that thirteen violations of daily máximums had occurred, had in essence established a prima facie case of thirteen daily violations. Requiring Gwaltney to come forward with some evidence to rebut this is not impermissible, particularly where, as here, such information would appear to be peculiarly [316]*316within the knowledge of Gwaltney. See Campbell v. United States, 365 U.S. 85, 96, 81 S.Ct. 421, 427, 5 L.Ed.2d 428 (1961) (“The ordinary rule, based on notions of fairness, does not place the burden upon a litigant of establishing facts peculiarly within the knowledge of his adversary.”) (quoting United States v. New York, N.H. & H.R.R., 355 U.S. 253, 256 n. 5, 78 S.Ct. 212, 214 n. 5, 2 L.Ed.2d 247 (1957)). Moreover, by requiring permit holders to monitor their discharges and report them through DMR’s, 33 U.S.C. § 1318(a)(3)(A); 40 C.F.R. § 122.41(1)(4), the Act itself places on the permit-holder a burden of showing compliance with the permit’s limitations. In addition, as we have stressed repeatedly, deciding the number of days of violation is relevant only in determining the maximum allowable penalty, and the district court is not obliged to assess the entire maximum amount. For these reasons, the court’s requiring Gwaltney to show that the actual number of days of violation was less than the thirteen violations that appeared on the face of its DMR’s, does not require a reassessment of the penalty imposed.
Gwaltney’s remaining contentions regarding the penalty assessed require little discussion.
As noted above, the district court and the parties agreed that in assessing penalties the court would use the EPA’s Civil Penalty Policy, 41 Env’t Rep. (BNA) 2991 (Feb. 16,1984), as a guideline. Gwalt-ney now complains that in adjusting the penalty to reflect the “gravity” of Gwalt-ney’s violations, the court exceeded the percentage increase recommended by the EPA policy. The EPA policy, however, is an internal document for agency guidance, and is not binding on the district court. The court agreed only to use the policy as a guideline, see 611 F.Supp. at 1556, and is not limited strictly by its terms in imposing an “appropriate” civil penalty. 33 U.S.C. § 1365(a).
Next, Gwaltney contends that the district court abused its discretion by assessing a penalty that failed to further the goal, espoused in the EPA Civil Penalty Policy, of “fair and equitable treatment of the regulated community.” Specifically, Gwaltney complains that the penalty assessed is much greater than those assessed in a handful of “similar” cases decided five or more years prior to this one. We do not undertake a detailed comparison of this and prior cases in order to determine the propriety of the penalty assessed below. Each case must stand on its own merits, and here the district court applied the agreed upon guidelines in a thorough, thoughtful, and rational manner. We do not view the outcome thus reached as arbitrary or an abuse of discretion.
Gwaltney also complains that the district court abused its discretion when, in adjusting the penalty to account for delays in compliance, it assessed a penalty for each day of unjustified delay, regardless of whether a violation actually occurred on each of those days. Because 33 U.S.C. § 1319(d) speaks in terms of penalties per day of violation, Gwaltney contends that the court erred in imposing penalties for days on which no violation occurred, and that its penalty should be reduced accordingly. We disagree. The Act speaks in terms of days of violation only for the purpose of establishing the maximum penalty assessable. Id. The district court properly considered Gwaltney’s days of violation in setting a maximum penalty of $6.66 million for 666 days of violation. Within that framework the court was permitted, in its discretion, to craft an “appropriate” penalty. Id. § 1365(a). This the court also did when, pursuant to the EPA Civil Penalty Policy, it penalized Gwaltney for unexcused delays in improving its effluent treatment systems. We see no abuse of discretion in the district court’s action.
III.
To summarize, for the reasons stated, we concur in the conclusion of the district court and hold that the absence of an ongoing violation does not defeat federal sub[317]*317ject matter jurisdiction over a citizen suit brought under section 505(a) of the Clean Water Act, and that civil penalties for past violations are an available remedy in such a suit.
We hold also that for the purpose of fixing an “appropriate” civil penalty which, pursuant to 33 U.S.C. § 1319(d), may be fixed at a maximum of $10,000 per day of violation, violations of “average” limitations encompassing periods greater than one day are to be treated as a violation for each day of the time period involved. The district court therefore properly employed this approach in setting Gwaltney’s maximum penalty, and did not abuse its discretion in calculating the penalty actually assessed against Gwaltney.
AFFIRMED.