United States v. Wilson , 133 F.3d 251 ( 1997 )

Reversed and remanded by published opinion. Judge NIEMEYER wrote the opinion for the court in parts I, II, V, and VI and Judge LUTTIG joined in parts I and V and Judge PAYNE joined in parts I, II, V, and VI. Judge LUTTIG wrote a separate opinion concurring in the judgment. Judge PAYNE wrote a separate opinion.

OPINION

NIEMEYER, Circuit Judge,

writing for the court on parts I, II, V and VI:

The defendants in this case were convicted of felony violations of the Clean Water Act for knowingly discharging fill and excavated material into wetlands of the United States without a permit. On this appeal they challenge: (1) the validity of federal regulations purporting to regulate activities that “could affect” interstate commerce; (2) the district court’s application of the Clean Water Act to wetlands that do not have a “direct or indirect surface connection to other waters of the United States”; (3) the district court’s application of the Clean Water Act to a practice known as “sidecasting” (depositing excavated material from wetland drainage ditches next to the ditch); (4) the district court’s interpretation of the mens rea required for a felony conviction under the Act; (5) evidentiary rulings of the district court; and (6) aspects of their sentences.

Because we conclude that 33 C.F.R. § 328.3(a)(3) (1993) (defining waters of the United States to include those waters whose degradation “could affect” interstate commerce) is unauthorized by the Clean Water Act as limited by the Commerce Clause and therefore is invalid, and that the district court erred in failing to require mens rea with respect to each element of an offense defined by the Act, we reverse and remand for a new trial.

I

In February 1996, after a seven-week trial, a jury convicted James J. Wilson, Interstate General Co., L.P., and St. Charles Associates, L.P., on four felony counts charging them with knowingly discharging fill material and excavated dirt into wetlands on four separate parcels without a permit, in violation of the Clean Water Act, 33 U.S.C. §§ 1319(c)(2)(A) & 1311(a). The district court sentenced Wilson to 21 months imprisonment and 1 year supervised release and fined him $1 million. It fined the other two defendants $3 million and placed them on 5 years probation. The court also ordered the defendants to implement a wetlands restoration and mitigation plan proposed by the government.

Wilson, a land developer with more than 30 years of experience, was the chief executive officer and chairman of the board of directors of Interstate General. He was personally responsible for various decisions relevant to the defendants’ convictions in this case. Interstate General was a publicly traded land development company with 340 employees, 2,000 shareholders, and assets of over $100 million. It was the general partner of St. Charles Associates, a limited partnership that owned the land being developed within the planned community of St. Charles, which lies between the Potomac River and the Chesapeake Bay in Charles County, Maryland. The convictions involve discharges onto four parcels that are part of St. Charles.

St. Charles currently consists of approximately 4,000 developed acres and 10,000 housing units with 33,000 residents. At completion, it is expected to be a 9,100 acre planned community of some 80,000 residents. The community was created under the New Communities Act of 1968 and developed initially in partnership between Interstate General and the United States Department of Housing and Urban Development (“HUD”). The project agreement provides for the creation of schools, parks, and recreational areas and designates at least 20% of the community to be reserved as “open space.” It also provides for the preservation of 75 acres of wetlands near Zekiah Swamp. In connection with the initial plan, HUD and Interstate General prepared an environmental impact statement, but that statement did not reflect any specific development plans for the four parcels involved in this case nor did it constitute a development permit under the Clean Water Act.

At trial, the government introduced evidence that during the period from 1988 to 1993, the defendants attempted to drain at least three of the four parcels of land, involved in this case by digging ditches. The excavated dirt was deposited next to the ditches — a process known as “sidecasting.” The government also introduced evidence that the defendants transported a substantial amount of fill dirt and gravel and deposited it on three of the parcels; only one parcel involved sidecasting without the addition of fill. The government presented evidence that all four of these parcels contained wetlands and that the defendants failed to obtain permits from the Army Corps of Engineers, the agency charged with issuing permits under the relevant section of the Act, 33 U.S.C. § 1344, prior to making efforts to drain and fill the parcels.

Although the parcels in question were not, because of neighboring development, located in pristine wilderness areas, the government presented substantial evidence about the physical characteristics which identified them as wetlands, including testimonial and photographic evidence of significant standing water, reports of vegetation typical to hydrologic soils, and infrared aerial photographs showing a pattern of stream courses visible under the vegetation. Evidence also showed that the properties were identified as containing wetlands on public documents including the National Wetlands Inventory Map and topographical maps of Charles County and the State of Maryland. The government demonstrated that water from these lands flowed in a drainage pattern through ditches, intermittent streams, and creeks, ultimately joining the Potomac River, a tributary of the Chesapeake Bay. ■

The government also produced evidence of the defendants’ awareness of the physical conditions of their land. The very development work underlying the present prosecution involved efforts to improve the drainage of the areas to make building feasible. Substantial fill was later added in an attempt to raise the ground level of the parcels. Some construction work involved repeated reshor-ing efforts because of wetness-induced ground shifting and collapse. Evidence was introduced that bids for work at one of the parcels actually contained different price quotations for wet and dry work because of the level of moisture on parts of the property. And witnesses gave testimony that despite the attempts at drying the property through ditching and draining or through the pumping off of standing water, and even after hundreds of truck loads of stone, gravel and other fill had been added to three of the parcels, wetland-loving plants continued to sprout through the fill.

Witnesses also testified at trial that a private consulting firm retained by the defendants informed the defendants that its observations of conditions on the parcels led it to conclude that the parcels contained wetlands. The firm recommended seeking permits from the Army Corps of Engineers before beginning development. The defendants were also contacted by Charles County zoning authorities concerned about the possible presence of wetlands in the vicinity of the new construction projects. Finally, the government presented evidence that even as the defendants complied with an Army Corps of Engineers order to cease construction on one of the parcels and remove fill dirt that had already been added, they continued to develop the other parcels without notifying the Corps or making an effort to ascertain whether a permit was necessary.

The defendants introduced contradictory evidence suggesting that whether the four parcels were wetlands under the Clean Water Act was unclear. They offered evidence which they claim showed that the Army Corps of Engineers was inconsistent in asserting jurisdiction over the parcels in question, claiming that the Corps took action on only one parcel, even though it had been aware for years of the ongoing development. Defendants also introduced an internal Corps memorandum that stated that while the areas in the St. Charles community have the “necessary parameters ... to be considered wetlands when using the Corps Wetland Delineation Manual,” “it is not clear to me that these areas can be interpreted as “waters of the United States’ within the meaning or purview of Section 404.” That memo suggested obtaining guidance from higher, authority as to what constitutes ‘“waters of the United States.” The defendants also introduced evidence indicating their belief that they had legally drained three parcels prior to introducing fill, and that no fill was discharged into the fourth which was being drained by the digging of ditches.

Following 15 hours of deliberation, the jury convicted all defendants of the four felony counts. Because of the felony convictions, the defendants were not convicted of four misdemeanor counts for the “negligent” violations of the Clean Water Act involving the same parcels.

II

The defendants challenge the authority of regulation 33 C.F.R. § 328.3(a)(3) (1993) (defining waters of the United States to include those waters whose degradation “could affect” interstate commerce) to extend jurisdiction of the Clean Water Act to the four parcels in question. They also challenge the district court’s instructions to the jury which were based on that regulation. They argue that the regulation and instructions exceed not only the authority of the Clean Water Act (which regulates “waters of the United States” without defining them), but also the Commerce Clause, U.S. Const, art. I, § 8, cl. 3. They maintain that in allowing the jury to find a nexus with interstate commerce based on whether activities “could affect” interstate commerce, the court authorized a “limitless view of federal jurisdiction,” far more expansive than the standard recently summarized in United States v. Lopez, 514 U.S. 549, 115 S.Ct. 1624, 131 L.Ed.2d 626 (1995).

While the defendants argue that the regulation and jury instructions are fatally flawed under Lopez because of their invocation of “potential” uses and effects on commerce, they do not challenge the constitutionality of the Clean Water Act itself.

In instructing the jury based upon 33 C.F.R. § 328.3(a)(3) (1993), the district court in this case said:

The government must prove that these waters have some potential connection with interstate commerce. If you find, ladies and gentlemen, beyond a reasonable doubt that these waters were or could be used by visitors from other states for recreational or other purposes, or that fish or shellfish are or could be taken from these waters and sold in interstate or foreign commerce, or that these waters were used or could have been used for industrial purposes by industries in interstate commerce, or that these waters were subject to the ebb and flow of the tide or that the use, degradation or construction [sic] [destruction?] of such waters could affect interstate commerce, then I instruct you as a matter of law that the government has established such connection with interstate commerce and that these waters, including wetlands, are waters of [the] United States.

The Clean Water Act prohibits the discharge, without a permit, of pollutants into “navigable waters.” 33 U.S.C. §§ 1311(a), 1362(12)(A). While the regulatory power of Congress over waters that are navigable in fact is well established, see, e.g., United States v. Rands, 389 U.S. 121, 122-23, 88 S.Ct. 265, 266, 19 L.Ed.2d 329 (1967) (“The Commerce Clause confers a unique position upon the Government in connection with navigable waters. ‘The power to regulate commerce comprehends the control for that purpose, and to the extent necessary, of all the navigable waters of the United States....’ ” (quoting Gilman v. Philadelphia, 70 U.S. (3 Wall.) 713, 724-25, 18 L.Ed. 96 (1865)), the Clean Water Act defines “navigable waters” as “the waters of the United States,” 33 U.S.C. § 1362(7). Construing the Clean Water Act, the Supreme Court has indicated that in defining “navigable waters” as “waters of the United States,” Congress intended “to exercise its powers under the Commerce Clause to regulate at least some waters that would not be deemed ‘navigable’ under the classical understanding of that term.” United States v. Riverside Bayview Homes, Inc., 474 U.S. 121, 133, 106 S.Ct. 455, 462, 88 L.Ed.2d 419 (1985). Accordingly, the power of Congress to regulate the discharge of pollutants into at least some nonnavigable waters is indisputable, but the limits of this power are far from clear. As explained in Lopez, Congress can clearly regulate discharges of pollutants that substantially affect interstate commerce. See 514 U.S. 549, 558-59, 115 S.Ct. 1624, 1629-30, 131 L.Ed.2d 626 (1995). Presumably, Congress may also regulate the discharge of pollutants into nonnavigable waters to the extent necessary to protect the use or potential use of navigable waters as channels or instrumentalities of interstate commerce, although the extent of that power is not entirely clear. Finally, it is arguable that Congress has the power to regulate the discharge of pollutants into any waters that themselves flow across state lines, or connect to waters that do so, regardless of whether such waters are navigable in fact, merely because of the interstate nature of such waters, although the existence of such a far reaching power could be drawn into question by the Court’s recent federalism jurisprudence. See, e.g., Printz v. United States, — U.S. -, 117 S.Ct. 2365, 138 L.Ed.2d 914 (1997); Seminole Tribe v. Florida, 517 U.S. 44, 116 S.Ct. 1114, 134 L.Ed.2d 252 (1996); Lopez, 514 U.S. 549, 115 S.Ct. 1624, 131 L.Ed.2d 626; New York v. United States, 505 U.S. 144, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992).

However, we need not resolve these difficult questions about the extent and limits of congressional power to regulate nonnavigable waters to resolve the issue before us. The regulation challenged here, 33 C.F.R. § 328.3(a)(3) (1993), defines “waters of the United States” to include:

All other waters such as intrastate lakes, rivers, streams (including intermittent streams), mud flats, sand flats, wetlands, sloughs* prairie potholes, wet meadows, playa lakes, or natural ponds, the use, degradation or destruction of which could affect interstate or foreign commerce----

(Emphasis added). This regulation purports to extend the coverage of the Clean Water Act to a variety of waters that are intrastate, nonnavigable, or both, solely on the basis that the use, degradation, or destruction of such waters could affect interstate commerce. The regulation requires neither that the regulated activity have a substantial effect on interstate commerce, nor that the covered waters have any sort of nexus with navigable, or even interstate, waters. Were this regulation a statute, duly enacted by Congress, it would present serious constitutional difficulties, because, at least at first blush, it would appear to exceed congressional authority under the Commerce Clause. This regulation is not, however, a statute. Absent a clear indication to the contrary, we should not lightly presume that merely by defining “navigable waters” as “the waters of the United States,” 33 U.S.C. § 1362(7), Congress authorized the Army Corps of Engineers to assert its jurisdiction in such a sweeping and constitutionally troubling manner. Even as a matter of statutory construction, one would expect that the phrase “waters of the United States” when used to define the phrase “navigable waters” refers to waters which, if not navigable in fact, are at least interstate or closely related to navigable or interstate waters. When viewed in light of its statutory authority, 33 C.F.R. § 328.3(a)(3) (1993), which defines “waters of the United States” to include intrastate waters that need have nothing to do with navigable or interstate waters, expands the statutory phrase “waters of the United States” beyond its definitional limit.

Accordingly, we believe that in promulgating 33 C.F.R. § 328.3(a)(3) (1993), the Army Corps of. Engineers exceeded its congressional authorization under the Clean Water Act, and that, for this reason, 33 C.F.R. § 328.3(a)(3) (1993) is invalid. For the same reason, the district court’s instruction based upon this regulation is also erroneous.

Ill

The defendants also contend that the district court, in instructing the jury on the relationship between wetlands and interstate waters, gratuitously and improperly extended the jurisdiction of the Clean Water Act beyond its stated limits to any wetland “even 'without a direct or indirect surface connection” with interstate waters. They claim that their property was not adjacent to waters of the United States and that any wetlands that may have been involved were too remote from navigable waters to be under the jurisdiction of the Clean Water Act. They observe that the wetlands involved here were “more than ten miles from the Chesapeake Bay, more than six miles from the Potomac River, and hundreds of yards from the nearest creeks.” They add that the district court’s interpretation of the nexus between wetlands and interstate waters also exceeds the Commerce Clause’s limitation of governmental power.

The government argues, without addressing the district court’s instruction, that the wetlands involved in this case “were [as a factual matter] clearly adjacent to streams which flow into the Chesapeake Bay” and therefore “were properly regulated pursuant to the Commerce Clause.”

Because of the divergence of these contentions and the presentation of conflicting evidence at trial, the question of whether the jury was properly instructed with respect to the scope of the Clean Water Act is material to whether the jury’s factual findings were properly informed by the law. Although the Clean Water Act itself authorizes the regulation of only “navigable waters,” which are defined as “waters of the United States,” see 33 U.S.C. §§ 1311(a), 1362(12)(A), 1362(7), the Army Corps of Engineers has defined “waters of the United States” to include wetlands adjacent to waters that otherwise constitute waters of the United States. 33 C.F.R. § 328.3(7) (1993). Although the Supreme Court has upheld the validity of 33 C.F.R. § 328.3(7) as a reasonable construction of the Clean Water Act, it did so explicitly in the context of a wetland “that actually abuts on a navigable waterway.” Riverside Bayview Homes, 474 U.S. at 135, 106 S.Ct. at 463 (emphasis added).

In instructing the jury in this case, the district court extended the application of the Clean Water Act substantially beyond the regulations that had been approved in Riverside Bayview Homes, instructing the jury that waters of the United States included adjacent wetlands “even mthout a direct or indirect surface connection to other waters of the United States.” (Emphasis added). This instruction intolerably stretches the ordinary meaning of the word “adjacent” and the phrase “waters of the United States” to include wetlands remote from any interstate or navigable waters. The magnitude of this extension is particularly highlighted when recognizing that the Army Corps of Engineers’ statutory mandate extends not to the regulation of “wetlands adjacent to waters of the United States,” but only to the regulation of “waters of the United States,” and that the Corps’ regulation of such wetlands is based solely on its definition of wetlands as “waters of the United States.” Furthermore, as noted above, we should interpret the Clean Water Act in light of the constitutional difficulties that would arise by extending the Act’s coverage to waters that are connected closely to neither interstate nor navigable waters, and which do not otherwise substantially affect interstate commerce. It was thus error for the district court to have instructed the jury to extend the jurisdiction of the Clean Water Act to wetlands that lack any “direct or indirect surface connection” to interstate waters, navigable waters, or interstate commerce.

Because we cannot determine as a result of this error whether the jury properly exercised the jurisdiction of the Clean Water Act, a new trial is necessary.

IV

The defendants next challenge the district court’s jury instruction that “sidecasting” in a wetland without the introduction of additional fill material to the wetland violates the Clean Water Act.

The government presented evidence that after the defendants attempted to drain the wetlands by digging ditches through them, but while the parcels retained sufficient hallmarks of wetland status to remain within the jurisdiction of the Army Corps of Engineers, large quantities of non-native fill materials were trucked onto the parcels in order to raise the ground level. The government claimed that the importation of that fill constituted the discharge of pollutants into the waters of the United States. The defendants did not deny that they added this fill. Rather, they attempted to show that at the time that fill was added, the ditching efforts had been successful in drying out the parcels, destroying their characteristics as waters of the United States, eliminating the risk of polluting'the now-absent water, and thus removing them from jurisdiction of the Clean Water Act. Resolving the factual question about the wetland-status of the parcels at the time the fill was added was clearly for the jury, and we do not undertake to resolve that dispute.

Included in the government’s claim that the defendants discharged pollutants in the waters of the United States, however, is the claim that the defendants’ very effort to drain the parcels involved a discharge of a pollutant into wetlands as prohibited by the Act. The government contended that in digging ditches in a wetland to drain it, the defendants deposited the excavated dirt next to the ditches and thereby discharged “a pollutant” into the wetland. The government acknowledges that if the defendants had removed the excavated dirt from the wetland, no violation would have resulted from the defendants’ ditching efforts, at least under the regulations then in effect, implicitly acknowledging that the anti-pollution enactment does not prohibit a non-polluting method of draining a wetland.

In instructing the jury, the district court stated:

The term discharge of dredged material and discharge of fill material ... inelude[s] the addition of fill material into wetlands for commercial and residential development, the grading or leveling of wetlands to remove water and to replace them with dry land or to change the bottom elevation, and the sidecasting of dredged material from the construction of drainage ditches in wetlands.

(Emphasis added). The court thus concluded that sidecasting violated the Clean Water Act. Therefore, the narrow question that the defendants present is whether the redeposit of wet soils excavated during the digging of a drainage ditch beside the excavation — i.e. “sidecasting” — constitutes a discharge of a pollutant in violation of the statute.

Recognizing the severe impact of the nation’s increasing water pollution on public health as well as on industries including fishing, agriculture, and outdoor recreation, Congress passed the Clean Water Act in 1972. The Act is a complex and comprehensive statute designed “to restore and maintain the chemical, physical, and biological integrity of the Nation’s waters.” 33 U.S.C. § 1251(a). Because pollution of the nation’s waters is the target of the Act, the Act prohibits the discharge, without a permit, of pollutants into navigable waters. 33 U.S.C. § 1311(a) & § 1362(12)(A). The term “discharge of a pollutant” is defined to mean “any addition of any pollutant to navigable waters from any point source.” 33 U.S.C. § 1362(12)(A) (emphasis added). Pollutant, in turn, is defined as “dredged spoil, solid waste, incinerator residue, sewage, garbage, sewage sludge, munitions, chemical waste, biological materials, radioactive materials, heat, wrecked or discarded equipment, rock, sand, cellar dirt and industrial, municipal, and agricultural waste discharged into water.” 33 U.S.C. § 1362(6).

While the district court instructed the jury somewhat expansively on the definition of a pollutant, including terms not included in the statute, such as “fill material, cement, and dredged materials resulting from excavation, ditching and land clearing operations,” we believe that its instructions in this regard were sufficiently consistent with the statutory definition of pollutant to enable the jury to conclude properly that the excavated dirt from drainage ditches was in fact a pollutant. Section 1362(6) includes as a pollutant “dredged spoil,” “biological materials,” “rock,” and “sand.” While the statutory term “dredged spoil” carries with it a more pejorative connotation than does the term that the court used, “dredged material,” the two are not sufficiently different to constitute error. Dictionary definitions of “spoil” include “material (as refuse earth or rock) excavated usually in mining, dredging, or excavating,” see Webster’s Third New International Dictionary 2203 (1961), or “earth and rock excavated or dredged,” Merriam Webster’s Collegiate Dictionary 1136 (10th ed.1994). Moreover, the statutory definition of pollutant also includes biological materials, rock, and sand which, the government argues, are the elements of the wet soil excavated from the wetland drainage ditches. Although the government must by proof establish that the dirt excavated from the wetland in fact falls within the statutory definition, we are satisfied that dredged materials, including the native soils excavated by ditching activities, may constitute a pollutant within the meaning of the Clean Water Act.

The question remains, however, whether “sidecasting,” which moves native wetland a few feet to the side of the ditch being created, constitutes a “discharge” within the meaning of the statute. The statute specifically states that “discharge” means “addition.” 33 U.S.C. § 1362(12). While the native soil is removed from the ditch and redeposited on the immediately adjacent land, the rational interpretation of the statute leads us to conclude that the movement of native soil a few feet within a wetland does not constitute the discharge of that soil into that wetland. The statute requires, in defining discharge of a pollutant, that the defendants have added dredged spoil to the wetland, the statutorily regulated water. While sidecasting moves excavated dirt from one particular locus in the wetland to another, it does not involve the addition of any material to the wetland. “Addition” requires the introduction of a new material into the area, or an increase in the amount of a type of material which is already present. While soil may be defi-nitionally transformed, through the act of excavation, from a part of the wetland into “dredged spoil,” a statutory pollutant, it is not added to the site. Were we to adopt so expansive a definition of “discharge” that any movement of soil within a wetland constitutes “addition,” we would not only flaunt the given definition of “discharge,” but we would be criminalizing every artificial disturbance of the bottom of any polluted harbor because the disturbance moved polluted material about. If Congress intended to reach such conduct, it need simply to redefine the term “discharge.” But as the statute is currently drafted, sidecasting does not involve the addition of pollutants to a water of the United States. But see Avoyelles Sportsmen’s League, Inc. v. Marsh, 715 F.2d 897, 923 (5th Cir.1983) (holding that “addition,” as included in the definition of “discharge,” could include “redeposit” of dredged material which need not come from an outside source).

In instructing the jury that sidecasting was prohibited by the Clean Water Act, we believe that the district court included conduct not prohibited by the Act and the regulations promulgated under it. Sidecasting from ditch-digging in itself effects no addition of a pollutant, and if the ditching successfully dries out the wetland prior to the addition of other materials, no violation of the Clean Water Act results because adding fill to dry-land cannot be construed to be polluting the waters of the United States. But if the ditching is unsuccessful in converting wetland to non-wetland and fill is added, a violation does result. In light of the evidence presented in this case, a jury may decide that the defendants’ conduct was culpable, but the error in the jury instructions requires a new trial.

V

The defendants also contend that the district court erred in instructing the jury about the criminal intent, the “mens rea,” required to prove a felony violation of the Clean Water Act. They argue (1) that the statute requires a showing that they were aware of the illegality of their conduct, and (2) that the required mens rea, however it is defined, must accompany each element of the offense. They note that the district court’s jury instructions comported with neither requirement.

The district court charged the jury that the government must prove each of four elements of the offense beyond a reasonable doubt:

First, that is the defendant knowingly ... discharged or caused to be discharged a pollutant.

Second, that the pollutant was [dis-jcharged from a point source.

Third, that the pollutant entered a water of the United States; and fourth, that the discharge was unpermitted.

The court defined an act as “knowingly” done “if it is done voluntarily and intentionally and not because of ignorance, mistake, accident or other innocent reason.” For each felony count, the court stated,

the government must prove that the defendants knew, one, that the areas which are the subject of these discharges had the general characteristics of wetland; and, two, the general nature of their acts. The government does not have to prove that the defendants knew the actual legal status of wetlands or the actual legal status of the materials discharged into the wetlands. The government does not have to prove that the defendants knew that they were violating the law when they committed their acts.

(Emphasis added). Finally, the court instructed on willful blindness, which it stated could stand in the place of actual knowledge.

Determining the mens rea requirement of a felony violation of the Clean Water Act requires us to make an interpretation based on “construction of the statute and ... inference of the intent of Congress.” United States v. Balint, 258 U.S. 250, 253, 42 S.Ct. 301, 302, 66 L.Ed. 604 (1922), quoted in Staples v. United States, 511 U.S. 600, 605, 114 S.Ct. 1793, 1797, 128 L.Ed.2d 608 (1994). We thus begin our analysis by looking at the language of 33 U.S.C. § 1319, as well as its place in the larger statutory structure.

Section 1319(e)(2)(A), making an illegal discharge of a pollutant a felony if accompanied by the defined mens rea, provides: “Any person who knowingly violates section 1311 ... shall be punished.” (Emphasis added). Section 1311 makes unlawful “the discharge of any pollutant” without a permit. And finally, § 1362 defines “discharge of a pollutant” to include “any addition of any pollutant to navigable waters from any point source” and defines “navigable waters” as “waters of the United States.” 33 U.S.C. §§ 1362(7) & (12). Within that statutory structure, we must determine the nature 'of intent that the statute requires for each element of the offense.

On a first reading of the clause, “any person who knowingly violates section 1311 shall be punished,” the order of words suggests that “knowingly” modifies “violates” so that the clause imposes punishment only when one violates the statute with knowledge that he is violating it, i.e. with knowledge of the illegality of his conduct. But the statute’s structure, the architecture of which includes a series of sections incorporating other sections, its legislative history, and the body of Supreme Court jurisprudence addressing mens rea of federal criminal statutes caution that our first reading may not so simply lead us to the proper interpretation.

Our first concern is a pragmatic one engendered by the overall structure of the Clean Water Act. The conduct that is made criminal with the “knowingly violates”' language encompasses numerous elements from other substantive statutory sections. See 33 U.S.C. § 1319(c)(2)(A). Each of those substantive sections may also be enforced with other civil and criminal penalties if the actions proscribed therein are performed with different scienter. See generally 33 U.S.C. § 1319. If Congress intended that the “knowing” mens rea accompany each element of the offense, as we have previously assumed is the case, see United States v. Ellen, 961 F.2d 462, 466 n. 2 (4th Cir.1992), the task of inserting the alternative mens rea requirements for the multiple civil and criminal enforcement provisions within each substantive prohibition would require confusingly repetitious drafting. A shorthand method of accomplishing the same purpose thus would be to insert “knowingly” in a single place where the conduct is made criminal, in this case, § 1319(c)(2)(A). See United States v. International Minerals & Chemical Corp., 402 U.S. 558, 562, 91 S.Ct. 1697, 1700, 29 L.Ed.2d 178 (1971) (the phrase “knowingly violates [applicable regulations]” was “a shorthand designation for specific acts or omissions which violate the act”).

Our second and more profound problem with our first-blush interpretative proposal arises from a recognition of two general common law principles regarding mens rea. First, in Anglo-American jurisprudence, criminal offenses are ordinarily required to have a mens rea. Staples, 511 U.S. at 605, 114 S.Ct. at 1796-97. This supposition is based on “the contention that an injury can amount to a crime only when inflicted by intention.” Id. (quoting Morissette v. United States, 342 U.S. 246, 250, 72 S.Ct. 240, 243, 96 L.Ed. 288 (1952)). Indeed, statutes requiring no mens rea are generally disfavored. See Staples, 511 U.S. at 606, 114 S.Ct. at 1797-98. But a second and deeply-rooted common law principle is that ignorance of the law provides no defense to its violation. See Cheek v. United States, 498 U.S. 192, 199, 111 S.Ct. 604, 609, 112 L.Ed.2d 617 (1991); Barlow v. United States, 32 U.S. (7 Pet.) 404, 411, 8 L.Ed. 728 (1833) (Story, J.). Thus, while some level of deliberateness is usually required to impose criminal punishment, it is also usually true that the defendant need not appreciate the illegality of his conduct. Applying those principles to a statute similar to the one before us, the Supreme Court in International Minerals declined “to attribute to Congress the inaccurate view that [the] Act requires proof of knowledge of the law, as well as the facts.” International Minerals, 402 U.S. at 563, 91 S.Ct. at 1701. In that case, the statute — which provided that whoever “knowingly violates any such regulation” shall be fined or imprisoned— was held to be a “shorthand designation” for knowledge of the specific acts or omissions which violate the Act. Id. at 559, 562, 91 S.Ct. at 1698-99, 1700. When so viewed, the Court noted, “the Act ... does not signal an exception to the rule that ignorance of the law is no excuse.” Id. at 562, 91 S.Ct. at 1700. Compare Liparota v. United States, 471 U.S. 419, 423 & 425 n. 9, 105 S.Ct. 2084, 2086-87 & 2088 n. 9, 85 L.Ed.2d 434 (1985) (holding that a statute providing “who[ ]ever knowingly uses, transfers, acquires, alters or possesses [food stamps] in a manner' not authorized by [the law]” was subject to a fine and imprisonment requires knowledge that the defendant know possession of food stamps was “unauthorized,” but stating that this interpretation creates no mistake-of-law defense). In light of these background rules of common law, we may conclude that mens rea requires not that a defendant know that his conduct was illegal, but only that he “know the facts that make his conduct illegal,” Staples, 511 U.S. at 605, 114 S.Ct. at 1797, unless Congress clearly specifies otherwise. And this knowledge must generally be proven with respect to each element of the offense. See, e.g., United States v. X-Citement Video, Inc., 513 U.S. 64, 78, 115 S.Ct. 464, 472, 130 L.Ed.2d 372 (1994) (construing a statute prohibiting knowingly shipping a visual depiction which involved “the use of a minor engaging in sexually explicit conduct” to require the defendant’s knowledge that the person depicted was in fact a minor); Ellen, 961 F.2d at 466 n. 2.

Finally, our first-blush reading of the phrase “knowingly violates” is cast into doubt by the legislative history, which suggests that Congress, by amending the statute in 1987, intended to facilitate enforcement of the Clean Water Act and increase the impact of sanctions by creating a separaté felony provision for deliberate, as distinct from negligent, activity. Before the amendment, the Act imposed a single set of criminal penalties for “willful or negligent” violations. See 33 U.S.C. § 1319(d)(1) (1986). The 1987 amendments, however, segregated the penalties for negligent violations, making them misdemeanors, and added felony provisions for knowing violations. See 33 U.S.C. § 1319(c)(1)(A) (“negligent” violation) & § 1319(c)(2)(A) (“knowing” violation). Thus, before 1987, the statute proscribed “willful or negligent” violations and after 1987 it proscribed separate “knowing” and “negligent” violations. In changing from “willful” to “knowing,” we should assume that Congress intended to effect a change in meaning. See United States v. Hopkins, 53 F.3d 533, 539 (2d Cir.1995); United States v. Sinskey, 119 F.3d 712, 716 (8th Cir.1997). Because “willful” generally connotes a conscious performance of bad acts with an appreciation of their illegality, see Ratzlaf v. United States, 510 U.S. 135, 141, 114 S.Ct. 655, 659, 126 L.Ed.2d 615 (1994), we can conclude that Congress intended to provide a different and lesser standard when it used the word “knowingly.” If we construe the word “knowingly” as requiring that the defendant must appreciate the illegality of his acts, we obliterate its distinction from the willfulness.

Based upon these interpretative guides, then, we cannot conclude that Congress intended to require the defendant to know that his eon duct was illegal when it stated that “Any person who knowingly violates [incorporated statutory sections] ... shall be punished.” The ready alternative interpretation is that Congress intended that the defendant have knowledge of each of the elements constituting the proscribed conduct even if he were unaware of their legal significance. This interpretation would not carry with it the corollary that the defendant’s ignorance of his conduct’s illegality provides him a defense, but would afford a defense for a mistake of fact. Thus, if a defendant thought he was discharging water when he was in fact discharging gasoline, he would not be guilty of knowingly violating the act which prohibits the discharge of pollutants. See United States v. Ahmad, 101 F.3d 386, 393 (5th Cir.1996); see also International Minerals, 402 U.S. at 563-64, 91 S.Ct. at 1700-01.

Accordingly, we hold that the Clean Water Act, 33 U.S.C. § 1319(c)(2)(A), requires the government to prove the defendant’s knowledge of facts meeting each essential element of the substantive offense, see Ellen, 961 F.2d at 466 n. 2, but need not prove that the defendant knew his conduct to be illegal, see International Minerals, 402 U.S. at 563, 91 S.Ct. at 1700-01.

Urging us to reach a different conclusion, the defendants argue that we should be governed by the decision in Liparota v. United States, 471 U.S. 419, 105 S.Ct. 2084, 85 L.Ed.2d 434 (1985), where the Court concluded that in proving a violation of a food stamp statute, the government had to prove that the defendant knew that his action or his possession of food stamps was unauthorized. The Court reached the conclusion there largely because, if it were not to require such a mens rea, the result would be to outlaw a number of acts which a reasonable person would very likely believe were entirely unregulated, including actions that were wholly accidental. The statute provided that “whoever knowingly uses, transfers, acquires, alters, or possesses[food stamps] in any manner not authorized by [the statute] or the regulations” was subject to a fine and imprisonment. 7 U.S.C. § 2024(b)(1). If the court did not apply a requirement that the defendant appreciate that his conduct was unauthorized, the result would render criminal “a nonrecipient of food stamps who ‘possessed’ stamps because he was mistakenly sent them through the mail due to an administrative error, ‘altered’ them by tearing them up, and ‘transferred’ them by throwing them away.” Id. at 426-27, 105 S.Ct. at 2089. The Court concluded that the evidence of Congress’ intent to create such a harsh regime was too meager to justify the interpretation. At the same time, however, the Court reiterated its conclusion that even then it did not provide a mistake-of-law defense, stating that no defendant could escape culpability by demonstrating that it did not know it to be illegal to use food stamps in an unauthorized way. Id. at 425 n. 9, 105 S.Ct. at 2088 n. 9.

In our case, we believe the several factors we have already discussed indicate a different congressional intent than that found by the Court in Liparota, as we too resist the temptation to create a mistake-of-law defense. As we noted, interpreting the phrase “knowingly violate” to mean violation with knowledge of an act’s illegality would require us to ignore the distinction between a knowing and a willful violation, a distinction that Congress recognized in amending the law in 1987. We also note that at the time the penalty provisions of the Clean Water Act were first written, the Supreme Court had already held in International Minerals that the use of the same construction at issue here — “knowingly violates [applicable regulations]” — was “a shorthand designation for specific acts or omissions which violate the Act.” 402 U.S. at 562, 91 S.Ct. at 1700. Thus, it is logical to conclude that Congress would have used a similar shorthand designation to incorporate the mens rea requirement against each of the substantive requirements found in the incorporated sections of the Clean Water Act without also intending to create a defense of ignorance of the law.

The defendants argue, however, that International Minerals has no bearing on this case because it was a case involving a “public welfare” offense. Under this somewhat amorphous exception to general common law scienter requirements, a threat to public health and safety posed by an object or activity and the inherent dangerousness or deleterious nature of the prohibited item are considered sufficient in themselves to place the defendant on notice of the likelihood of regulation and thus to excuse the need to prove mens rea with respect to one or more elements of the offense. Even under this public welfare doctrine, however, true or rigid strict liability does not generally follow, as ignorance of the facts usually remains a defense. See Staples, 511 U.S. at 607 n. 3, 114 S.Ct. at 1798. Thus, International Minerals held that prosecution for transporting sulfuric acid without following proper regulatory procedures under a statute prohibiting one from “knowingly violating] any such regulation” did not establish an ignorance of-the-law defense by requiring the government to prove the defendant knew of the regulation and appreciated that his conduct violated it. 402 U.S. at 563-64, 91 S.Ct. at 1700-01. Rather, the government would have to prove that the defendant knew of the operative facts which themselves were the essential elements of the regulatory violation. Id. The Court thus preserved mistake-of-fact defenses and did not, at the same time, create true strict liability. Later cases which did not involve “public welfare” offenses have required no more than knowledge of the facts. See, e.g., Staples, 511 U.S. at 606, 114 S.Ct. at 1797-98.

The fact that International Minerals involved regulations of an inherently deleterious substance of a type not involved in the present prosecution does not undercut our belief that Congress did not here intend to create a mistake-of-law defense. Even though the materials involved in this case, fill and native soil from a wetland, may not be inherently deleterious, the Clean Water Act is, as a general matter, largely concerned with pollutants that are inherently deleterious. The legislative history of the Act records Congress’ explicit concern with public health. See S.Rep. No. 92-414 (1972), reprinted in 1972 U.S.C.C.A.N. 3668, 3670-71. And the Act specifically authorized research to determine the harmful effects of pollutants on the health or welfare of persons. See 33 U.S.C. § 1254(c). To date, three other circuits have concluded that the Clean Water Act involves public welfare offenses, recognizing that in each of those cases the pollutants were inherently deleterious. See Sinskey, 119 F.3d at 716 (meat packing plant waste water containing large amounts of ammonia nitrate); Hopkins, 53 F.3d at 534 (waste water containing zinc and other toxic chemicals); United States v. Weitzenhoff, 35 F.3d 1275, 1284 (9th Cir.1998) (sewage).

While a statute which in some applications is a public welfare statute may in other applications be held to require a different mens rea, see Staples, 511 U.S. at 605, 114 S.Ct. at 1796-97, even in the latter situation, the government need prove only that the defendant knew the operative facts which make his action illegal. The government need not prove that the defendants understood the legal consequences of those facts or were even aware of the existence of the law granting them significance. Compare United States v. Freed, 401 U.S. 601, 609, 91 S.Ct. 1112, 1118, 28 L.Ed.2d 356 (1971) (holding that violation for possession of unregistered hand grenade under 26 U.S.C. § 5861(d) was public welfare offense and the government was not required to prove that defendant knew grenade to be unregistered) with Staples, 511 U.S. at 619, 114 S.Ct. at 1804 (holding that violation for possession of unregistered machine gun under 26 U.S.C. § 5861(d) was not a public welfare offense and the government thus was required to prove defendant’s knowledge of the specific charae-teristics which qualify the weapon as a statutory machine gun); see also Hamling v. United States, 418 U.S. 87, 119-121, 94 S.Ct. 2887, 2908-10, 41 L.Ed.2d 590 (1974) (requiring government to prove that a defendant knew the contents of a mailed package, but not their legal status as obscene materials); accord Rosen v. United States, 161 U.S. 29, 32-33, 16 S.Ct. 434, 435 (1896).

In light of our conclusion that the government need only prove the defendant’s knowledge of the facts meeting each essential element of the substantive offense and not the fact that defendant knew his conduct to be illegal, in order to establish a felony violation of the Clean Water Act, we hold that it must prove: (1) that the defendant knew that he was discharging a substance, eliminating a prosecution for accidental discharges; (2) that the defendant correctly identified the substance he was discharging, not mistaking it for a different, unprohibited substance; (3) that the defendant knew the method or instrumentality used to discharge the pollutants; (4) that the defendant knew the physical characteristics of the property into which the pollutant was discharged that identify it as a wetland, such as the presence of water and water-loving vegetation; (5) that the defendant was aware of the facts establishing the required link between the wetland and waters of the United States;^* and (6) that the defendant knew he did not have a permit. This last requirement does not require the government to show that the defendant knew that permits were available or required. Rather, it, like the other requirements, preserves the availability of a mistake of fact offense if the defendant has something he mistakenly believed to be a permit to make the discharges for which he is being prosecuted.

While we thus reject the defendants’ challenge to the district court’s instructions based on the contention that the government must prove awareness of the illegality of them conduct, we agree that the instructions did not adequately impose on the government the burden of proving knowledge with regard to each statutory element. For this reason, a new trial is required.

VI

Finally, defendants challenge the district court’s rulings that allowed into evidence expert opinions offered by the government on the proper interpretation of the applicable law but that excluded the testimony of two former Assistant Attorneys General of the United States by which the defendants proposed to establish their understanding of the proper interpretation of the law. Because we have already identified reasons for requiring a new trial, we need not resolve whether the court’s ruling constitutes reversible error. Only because we believe it will be helpful in connection with a new trial, and might prevent further appeals, do we address these evidentiary rulings.

In a case involving the application of complex regulations, like this one, it can be difficult for factual witnesses who are also the individuals responsible for enforcing the regulations to testify without at least referencing the legal framework which motivated and justified their activities. It can likewise be difficult for the trial court to preserve to itself the exclusive task of instructing the jury on the law while, at the same time, allowing expert witnesses to state their findings and to give their opinions on relevant questions. Nonetheless, the jury must be instructed on the law by the court and not by the witnesses. As we have explained in a similar circumstance:

[U]nder our system it is the responsibility—and the duty—of the court to state to the jury the meaning and applicability of the appropriate law, leaving to the jury the task of determining the facts which may or may not bring the challenged conduct within the scope of the court’s instruction as to the law.

Under circumstances involving domestic law, this court can conceive of no circumstances which would shift this burden from the court to the jury, where the jury judgment would be influenced, if not made, on the basis of expert witness testimony which would undoubtedly follow the usual pattern of conflicting expert opinions. Permitting such testimony as to legal conclusions gives cogent meaning to the “apprehensions that jurors will turn to the expert, rather than to the judge, for guidance on the applicable law.”

* * *

The proffer of expert opinion in many cases raises problems difficult of resolution by the trial court, where the line must be drawn between proper expert evidence as to facts, the inferences to be drawn from those facts, and the opinions of the expert, on the one hand, and testimony as to the meaning and applicability of the appropriate law, on.the other hand. While sometimes difficult to discern that line, especially in the heat of trial, it nonetheless must be drawn.

Adalman v. Baker, Watts & Co., 807 F.2d 359, 366 (4th Cir.1986) (citations omitted).

We thus conclude that the district court acted well within its discretion in excluding the proffered testimony of the two former Assistant Attorneys General. With regard to its rulings on the testimony of government expert witnesses, the issue is more difficult, because those witnesses were also fact witnesses to the conditions of the property in question and expert witnesses as to techniques for identifying wetlands. As regulatory enforcers, their conclusions that a given property contained wetlands were based not purely upon the scientific notion of what a wetland is, but also upon the regulatory definition. Nevertheless, we believe that to maintain properly the court’s role as the sole arbiter of the applicable law, the court should have taken steps to limit those witnesses’ testimony generally to facts of history, practices and procedures followed by them in their work, opinions based on demonstrated expertise, and similar matters, but it should not have allowed them to give opinions on what the law means or how it is interpreted.

VII

In summary for the reasons given in parts II and V, we conclude (1) that 33 C.F.R. § 328(a)(3) (defining, in part, “waters of the United States”) exceeds statutory authority and therefore is invalid; and (2) that the district court erred in failing to apply the statutory mens rea to each element of the offense. For these reasons, we order a new trial. Because we are ordering a new trial, we need not reach the defendants’ sentencing issue.

REVERSED AND REMANDED FOR A NEW TRIAL.

* These facts might seem to fall into the category of "jurisdictional facts” which under United States v. Feola, 420 U.S. 671, 676 n. 9, 684, 95 S.Ct. 1255, 1260 n. 9, 1263-64, 43 L.Ed.2d 541 (1975), the government need not prove the defendant knew. In Feola, however, those facts served only the purpose of designating which of two authorities would prosecute the crime, the underlying crime of assault being criminal at all times either under state law or, if involving a federal official, as in that case, under federal law. See Feola, 420 U.S. at 683, 95 S.Ct. at 1263. Maryland does, by statute, regulate the discharge of pollutants into certain waters within its jurisdiction, as well as prohibiting the destruction of wetlands, see Md.Code Ann., Envir. §§ 4-101 to 4-708; § 16-302, but as those statutes do not appear to cover the property here in question, we do not find this situation to be governed by Feola.