United States Court of Appeals
                              FOR THE EIGHTH CIRCUIT
                                    ___________

                                    No. 00-2430
                                    ___________

United States of America,                *
                                         *
      Plaintiff - Appellee,              *
                                         * Appeal from the United States
      v.                                 * District Court for the
                                         * Southern District of Iowa.
Dico, Inc.,                              *
                                         *
      Defendant - Appellant.             *
                                    ___________

                               Submitted: April 10, 2001

                                   Filed: September 19, 2001
                                    ___________

Before BOWMAN and FAGG, Circuit Judges, and PIERSOL,1 District Judge.
                           ___________

BOWMAN, Circuit Judge.

      The United States sued Dico, Inc., pursuant to the Comprehensive
Environmental Response, Compensation, and Liability Act (CERCLA), 42 U.S.C.
§§ 9601-9675 (1994 & Supp. IV 1998), to recover response costs incurred for the




      1
        The Honorable Lawrence L. Piersol, Chief Judge, United States District Court
for the District of South Dakota, sitting by designation.
cleanup of groundwater contaminated by trichloroethylene (TCE)2 and other
chlorinated volatile organic compounds (VOCs). The District Court3 held a bench trial
and found Dico liable for response costs incurred in connection with the groundwater
contamination. The court also granted summary judgment to the United States on the
amount of cleanup costs it was entitled to recover from Dico. Dico appeals both as to
liability and damages. We affirm.

                                           I.

       This litigation returns to this Court for the third time. The case arises from the
EPA's determination, made in the mid-1970s, that the Des Moines public water supply
had been contaminated by TCE and other chlorinated VOCs. Dico's property sits
within the boundaries of the Des Moines TCE Site (the Site), the land area identified
by the EPA as the source of the contamination. Dico and other businesses located
within the Site had for many years used TCE for degreasing and other industrial
applications. The EPA designated Dico as a potentially responsible party under
CERCLA and issued a cleanup order to Dico. Dico complied, thereby incurring
response costs. Moreover, the EPA itself incurred costs in connection with the cleanup
of the Site. For a detailed history and factual background of the prior litigation, see
Dico 

(reversing grant of summary judgment to the EPA in Dico's
suit against the EPA seeking reimbursement for response costs it incurred in complying
with the EPA's cleanup order), and United States v. Dico, Inc., 

, 574-75
(8th Cir. 1998) (Dico II) (reversing grant of summary judgment to the EPA in the EPA's
cleanup-cost recovery action against Dico).


      2
       "TCE is a suspected human carcinogen and also has been linked to neurological
damage, and, at high exposure, death." Dico, Inc. v. Diamond, 

, 349 n.2
(8th Cir. 1994) (Dico I).
      3
      The Honorable Ronald E. Longstaff, United States District Judge for the
Southern District of Iowa.
                                           -2-
       Following our remand in Dico II, the District Court heard the EPA's suit against
Dico at a bench trial. The EPA's theory of the case was that Dico had released TCE
on its property via numerous sources and activities: leaks from vapor degreasers that
used TCE to degrease metal parts manufactured by Dico; spreading TCE-laden sludge
as a dust suppressant on the ground around Dico's buildings; dumping TCE sludge
directly onto the ground; storing TCE sludge in drums left exposed to the weather;
leaks from railcars, drums, and storage tanks used in Dico's chemical repackaging and
supply business; and cleaning of TCE storage drums at Dico's facility. The EPA further
contended that the TCE released by Dico migrated through the soil into the
groundwater below Dico's property.

       In defense, Dico attempted to identify potential alternative sources for the
groundwater contamination. Dico argued that all of the groundwater contamination the
EPA attributed to its activities actually originated with polluters located to the north of
its property. Dico contended that contamination found a mile north of Dico, near the
Martin Luther King Expressway (MLKE area), migrated with groundwater flowing
south into the aquifer below Dico's property. Dico also attacked the government’s case
by arguing that even if Dico released TCE, none of the TCE could have migrated far
enough down through the soil to reach the groundwater.

       The District Court found Dico liable for the costs incurred by the EPA in
cleaning up the groundwater at the Site, but delayed its decision on the amount of the
cleanup costs to be awarded to the EPA pending a hearing on the EPA's motion for
summary judgment on that question. After a hearing, the District Court granted
summary judgment to the EPA, awarding it $4,129,426.67 in cleanup costs. Dico
appeals both the liability and the award portions of the judgment entered by the District
Court.




                                            -3-
                                          II.

       Dico challenges the District Court's liability determination on numerous grounds.
We first address Dico's objections to evidentiary rulings on the admission of expert
testimony and of deposition testimony offered by the government after the close of both
parties' cases-in-chief.

                                          A.

       Dico argues that the District Court erred in refusing to exclude the testimony of
John Robertson, the government's expert hydrogeologist, because "his methodology
was unreliable." Appellant's Br. at 29. We review a district court's decision to admit
expert testimony under Daubert v. Merrell Dow Pharmaceuticals, Inc., 

(1993), and Kumho Tire Co. v. Carmichael, 

(1999), for an abuse of
discretion. 

; General Elec. Co. v. Joiner, 

, 138-39
(1997). Admissible expert testimony must be grounded upon scientifically valid
reasoning or methodology.4 

. The court must examine
both the relevance and the reliability of the proffered testimony, Blue Dane Simmental
Corp. v. Am. Simmental Ass'n, 

, 1040 (8th Cir. 1999), keeping in mind
that “[t]he focus, of course, must be solely on principles and methodology, not on the
conclusions that they generate.” 

.

      Dico does not dispute the relevance of Robertson's testimony. Rather, Dico
challenges the reliability of Robertson's testimony, urging us to reject his methodology

      4
       Some of the factors relevant to this inquiry include whether the theory or
technique at issue can be tested, whether it has been subjected to peer review and
publication, whether its rate of error is known, whether there are standards that are
maintained to control the technique's operation, and whether the technique has gained
acceptance within the relevant scientific community. 

.
                                          -4-
and exclude his testimony on several grounds. First, Dico argues that Robertson's
analysis of the contamination found on its property is fundamentally flawed because he
ignored evidence suggesting alternative sources for the contamination. Dico argues that
Robertson improperly excluded from his computer model any consideration of
groundwater flow onto Dico's property from the north,5 failed to consider the Ingersoll
Run as a conduit for contamination originating off of Dico's property,6 disregarded data
indicating that the amount of TCE in the Dico remediation wells increased after their
activation,7 and ignored all data regarding TCE contamination in the MLKE area.


      5
        Dico also argues that Robertson mischaracterized the soil permeability in the
north area as low because the United States Department of Agriculture (USDA)
classifications of soil in this area indicated moderate permeability. Robertson testified
that the USDA classifications apply only to the upper three to six feet of the soil, and
thus are primarily meant for agricultural purposes. Tr. at 152. He distinguished these
classes from those of the deeper soils in the area, and noted that the wells drilled there
could not produce water as would wells drilled in an aquifer. Thus, he reasoned that
the permeability of the soil was low compared to the surrounding areas. See Tr. at 129-
30.

        Moreover, Dico urges us to reject Robertson's analysis because the EPA's report
on the north area stated that a permeable sand lens thirty feet thick underlaid parts of
that location. See Tr. at 342. Dico's argument is unpersuasive; the report itself further
states that the "sand lens does not appear to be connected hydraulically to the lowland
flood plain alluvial deposits" found below the Racoon River. U.S. EPA, Remedial
Investigation Report, Des Moines TCE OU NO. 3, 1-3 (June 24, 1992). The Racoon
River sits between Dico and the north area, and thus the sand lens appears not to factor
into groundwater flow toward Dico from the north. The District Court did not abuse
its discretion in rejecting Dico's arguments on these issues.
      6
       The Ingersoll Run is a drainage ditch that runs from north to south through the
eastern portion of Dico's facility and terminates in a drainage pond at the south end of
the property.
      7
       Dico argues that this data supports its defense that the pollution that was
recovered actually originated elsewhere and was sucked into the aquifer below Dico's
                                           -5-
       The record undermines Dico's characterization of Robertson’s analysis. First,
Dico’s arguments regarding Robertson’s computer model are inapposite because
Robertson testified that the model did not form the basis for his conclusion regarding
the origin of the contamination. Tr. at 156. Robertson used the model for a limited
purpose—studying the capture zones of the remediation wells within the boundaries of
Dico’s property. Tr. at 157. Thus, Dico’s complaints regarding the data Robertson put
into the model are without consequence to the validity of his analysis of alternative
sources for the contamination. Furthermore, the model itself passes scrutiny under
Daubert. Known as MODFLOW, the model is sanctioned by the EPA and is
considered a standard model that is acceptable and commonly used by hydrogeologists.
Tr. at 306. Dico's expert hydrogeologist used the same computer model as Robertson
in his evaluation of the Site. 

also shows that he considered each piece of data that Dico
alleges he ignored. He looked at data regarding whether the Ingersoll Run was a
conduit of contamination onto Dico's property. Citing the negative results of tests
performed on soil borings taken close to the Run and on sediment samples from the
pond at its terminus, Robertson rejected the Run as a possible source. Tr. at 73-74.
Moreover, Robertson's testimony reflects that he considered the data indicating that the
amount of TCE in the Dico remediation wells increased after their activation. He
explained that he used the continued existence of contamination to support his
conclusion that TCE in dense non-aqueous phase liquid form (DNAPL) existed in the
soil beneath Dico's property. Tr. at 70-71. Similarly, Robertson did not ignore data
regarding contamination in the MLKE area. Rather, he testified that he examined data
in the reports on the area and considered the amounts and location of the contamination
in relation to the topography between the MLKE area and Dico's facility. Tr. at 128.
From his study, Robertson concluded that the clearly delineated borders of the
contamination and the different chemical fingerprints of the chemicals of concern


property by the operation of the wells.
                                          -6-
present at the MLKE area precluded that site from being the source of the TCE
removed from the groundwater by the Dico remediation well system. Dico’s arguments
regarding the reliability of Robertson’s analysis of alternative sources of contamination
do not show an abuse of discretion by the District Court in admitting Robertson’s
testimony.

       Dico argues, as a second ground for excluding Robertson’s testimony, that his
analysis of the migration of TCE from the soil into the groundwater at Dico's property
is based on unreliable methodology. Dico claims that Robertson’s "continuous line
theory" and conclusion that the contaminants left a trail of DNAPL in the soil are
unsupported by the record. Dico's challenge rests primarily on the argument that the
number of soil samples taken was insufficient to support Roberton's conclusions. As
part of its examination of the contamination on Dico's property, Dico's contractor,
Eckenfelder, Inc., drilled sixty-nine deep soil borings on Dico's property, each
approximately twenty-five feet deep. From each twenty-five foot boring, approximately
two small samples were sent to a laboratory for analysis. Dico contends that even if
every sample tested contained TCE, such a small section of each boring was examined
that the results cannot reliably support the conclusion that contamination existed
continuously from the surface to the groundwater.

       Robertson specifically testified that he did not rely on these results to conclude
that a continuous line of contamination existed; rather, he used the results as a check
that bolstered a conclusion he had already reached based upon other evidence
collected. Because Dico challenges Robertson's conclusion by citing evidence that he
did not use to form that conclusion, Dico's challenge fails. Moreover, the sufficiency
of the factual basis of Robertson's continuous line theory was open to any challenge
Dico desired to mount on cross-examination, but that sufficiency was not a basis for
excluding Robertson's testimony altogether. See Hose v. Chicago N.W. Transp. Co.,

, 974 (8th Cir. 1995) (stating that sufficiency of factual basis of expert
testimony goes to credibility, not admissibility, unless expert's opinion "is 'so

                                           -7-
fundamentally unsupported that it can offer no assistance to the jury'" (quoting
Loudermill v. Dow Chem. Co., 

, 570 (8th Cir. 1988))). Robertson cited
several sources of evidence for his opinion: the organic vapor analyzer readings
indicating the presence of chlorinated VOCs in the soil, the composition of the
chemicals present in the soil, and the long-term presence of contamination despite the
implementation of the remediation system. The District Court did not abuse its
discretion in concluding that Robertson offered sufficient factual support for his opinion
to withstand Dico's challenge to its admissibility.

      As to Dico's attack upon Robertson's conclusions regarding the presence of
DNAPL in the soil at the Dico facility, Robertson postulated that the existence of
DNAPL in the soil accounts for the continued recovery of TCE by the remediation
wells on Dico's property, in contrast to the defense theory that the remediation wells
sucked contaminated groundwater from the MLKE area into the aquifer below Dico's
property. Robertson testified that Eckenfelder based its conclusion that DNAPL was
not present in the soil on a solubility threshold that was ten times greater than the
standard accepted in the field. Several of the measurements of TCE taken by
Eckenfelder for its report met the currently accepted standard and thus indicated the
presence of DNAPL in the soil. Therefore, his opinion was not contrary to the
published data and reports; rather, it used that data and applied a different and
professionally accepted standard. Robertson's conclusion is, moreover, in accord with
the conclusion drawn by Dico's consultant, Eckenfelder, that DNAPL was probably
present in the soil at Dico's facility. See, e.g., Tr. at 147. The District Court did not
abuse its discretion by refusing to preclude Robertson's testimony on this basis.

       Third, Dico asserts that Robertson’s testimony is unreliable because his findings
are, in at least three instances, based upon insufficient data. Again, because "the
factual basis of an expert opinion goes to the credibility of the testimony, not the
admissibility, and it is up to the opposing party to examine the factual basis for the
opinion in cross-examination," the District Court properly refused Dico's invitation to

                                           -8-
exclude Robertson's testimony on this ground. Loudermill v. Dow Chem. Co., 

, 570 (8th Cir. 1988) (holding district court did not abuse its discretion by
admitting expert testimony challenged by defendant as being "insufficiently supported
by facts"); accord Arkwright Mut. Ins. Co. v. Gwinner Oil, Inc., 

, 1183
(8th Cir. 1997) (same).

       In sum, Dico's objections to the District Court's decision to admit Robertson's
expert testimony amount to an argument that the District Court should have given more
weight to Dico’s expert’s interpretation of the data at issue.8 Dico has not pointed to
any deficiency in the reliability of Robertson’s testimony that would lead us to conclude
the District Court abused its discretion.

                                           B.

      Hydrogeologist Abdul S. Abdul provided expert testimony on behalf of Dico.
The District Court excluded several opinions offered by Abdul because they were not
contained in his expert report or otherwise disclosed to the United States prior to the
bench trial. Dico argues that the District Court erred.

      Dico's objection to the District Court's ruling on Abdul's opinion testimony is
without consequence. The court specifically stated that "the outcome of the litigation
would not change even if the above opinions were admitted due to the fact [that] Dr.


      8
       Dico’s argument that the District Court misapplied the burden of proof on the
admissibility of Robertson’s testimony is without merit. The record shows that the
government laid an adequate foundation for the admission of his testimony. "Vigorous
cross-examination [and] presentation of contrary evidence . . . are the traditional and
appropriate means of attacking shaky but admissible evidence." 

. By noting that Dico failed to present evidence of any other methodology that
Robertson should have applied, the court merely emphasized that Dico's Daubert
challenge was not, in the court's view, well-founded.
                                           -9-
Abdul and Dico has [sic] failed to establish a clear connection between the
contamination found in the MLKE project area and that pulled from the groundwater
beneath the Dico property." United States v. Dico, Inc., No. 4-95-10289, at 18 n.12
(S.D. Iowa Feb. 14, 2000) (Ruling on Liability and Order Setting Hearing Date).
Furthermore, the excluded opinions, even if considered, would not persuade this court
that the District Court erred in its ruling on the admission of Robertson's testimony (or,
for that matter, in its findings on causation, see infra pt. III). 

pt. II.A.

                                           C.

       Dico argues that the District Court abused its discretion in admitting deposition
testimony from three witnesses that the government offered into evidence after both
sides had rested their cases. Dico had filed pretrial motions to exclude these
depositions, but the District Court had ruled that the deposition testimony was
admissible. The government did not specifically offer the depositions during its case-
in-chief. After Dico finished presenting its case, the District Court asked the
government's counsel about the depositions. The government responded that it had
intended to offer the deposition testimony, but that it had apparently misunderstood the
court's procedure regarding when and how that testimony would be admitted. The
government indicated that it would like to have the testimony admitted and, over Dico's
objections, the court admitted the deposition testimony into evidence after the close of
Dico's case.

       It has been our long-standing judgment that "[a] court has, of course, the general
power to reopen a case, either on motion of a party or on its own motion, while the
matter is still under advisement, for the receipt of further evidence." Arthur Murray,
Inc. v. Oliver, 

, 34 (8th Cir. 1966). Nonetheless, "it is not the business of
a court to make additions of evidence in a submitted case on its own motion other than
as there may be some element of such probative importance that its addition will


                                           -10-
prevent a miscarriage of justice from occurring in the situation." 

the
exercise of such power for an abuse of discretion. See 

that the District Court did not abuse its discretion by admitting the
disputed deposition testimony offered by the government. At the time the court decided
to admit the disputed testimony, the court offered Dico the opportunity to present
additional evidence in response to it. See Briscoe v. Fred's Dollar Store, Inc., 

, 1028 (8th Cir. 1994) ("It is not an abuse of discretion for a district court to
reopen a plaintiff's case where a defendant is given an opportunity to challenge the new
evidence introduced."); see also Ingram v. Mo. Pac. R.R. Co., 

, 1455 n.5
(8th Cir. 1990) (finding no abuse of discretion where plaintiff was permitted to
introduce evidence after both sides had rested). Moreover, Dico was not surprised by
the government's evidence; it had been the subject of pretrial motions and a pretrial
order, and Dico had submitted counterdesignations of deposition testimony to be
admitted in response to the sections offered by the government. The court
unambiguously assured Dico that it would consider both the government's designations
and Dico's counterdesignations. Finally, the court allowed Dico the opportunity to
review the testimony as it was marked and submitted in the court's bench book to
further ensure that Dico was satisfied with the manner in which the testimony and
Dico's counterdesignations were submitted. The District Court exercised great caution
in affording Dico an adequate opportunity to respond to the testimony at issue. We see
no abuse of discretion in the District Court’s actions.

                                          III.

       Dico next challenges the judgment against it by arguing that the government
failed to meet its burden of proof on causation. Dico attacks the two factual findings
upon which the District Court's liability holding rests. The District Court found that
releases of hazardous substances occurred on Dico's property and that a causal nexus
exists between these releases and the groundwater contamination at the Site. As part

                                          -11-
of the second finding, the court found that the hazardous materials released on Dico's
property migrated through the soil to reach and contaminate the groundwater. We
review these findings of fact for clear error. Control Data Corp. v. S.C.S.C. Corp., 

, 938 (8th Cir. 1995).

                                          A.

        CERCLA defines “release” to mean “any spilling, leaking, pumping, pouring,
emitting, emptying, discharging, injecting, escaping, leaching, dumping, or disposing”
of hazardous wastes “into the environment.” 42 U.S.C. § 9601(22) (emphasis added).
The District Court found that TCE was released into the soil on Dico's property in four
ways. First, "some quantity of TCE leaked from known cracks in the degreasing vat
through the concrete containment pit and into the ground underneath." United States
v. Dico, Inc., No. 4-95-10289, at 10 (S.D. Iowa Feb. 14, 2000) (Ruling on Liability and
Order Setting Hearing Date). Second, "Dico routinely dumped and spread sludge
containing TCE residue directly on its property." 

quantities of TCE
leaked during the filling of storage drums with TCE. 

delivering
TCE to Dico dripped the chemical onto the soil while on Dico's property. 

in the record supports the District Court's finding that Dico released TCE and
that it reached the soil.

                                    1. TCE Leaks

       The government presented evidence of numerous TCE leaks. William O'Brien,
a former Dico employee, stated in his deposition testimony that in 1975 he discovered
a significant leak in the degreaser. The degreaser was using an unusually large amount
of TCE, and when O'Brien had the degreaser pulled out of its concrete containment pit
for examination, he observed cracks in the bottom of the degreaser. Despite these
cracks, he found no TCE pooled in the containment pit. Dep. of William J. O'Brien,
July 12, 1995, at 56-57, 136-37. O'Brien described the absence of pooled TCE as an

                                         -12-
"amazing thing." 

John Harold Strouf, a retired president of Dico, also
testified to leaks in the degreaser. Strouf confirmed that Dico used at least two
degreasers, one installed above ground and one installed in a concrete pit. Dep. of John
Harold Strouf, Sept. 30, 1996, at 37. He testified that leaks from the degreaser tanks
were caused by holes torn in the sides during the raising and lowering of the baskets
containing parts to be degreased. 

        Moreover, O'Brien testified that the company disposed of sludge containing
TCE, collected from the bottom of the degreaser, by dumping the sludge on the ground
about 100 feet from the production building and covering it with fill material and dirt.
Dep. of William J. O'Brien at 27; see also 

(during 1950s TCE waste sludge
from degreaser was dumped on property but in different location). Strouf testified that
Dico took the TCE sludge from the bottom of the degreaser and used it as a dust
suppressant on roadways around the Dico facility. He stated that this practice was
already established before he started working at Dico in 1960, and that as many as two
fifty-five gallon barrels of sludge were removed from the company's degreasers each
time they were cleaned. Dep. of John Harold Strouf at 44-45.

       Finally, the record contains further evidence of leaks resulting from the chemical
supply business that Dico ran from 1946 until 1980. TCE arrived at Dico's facility in
railcars, from which Dico pumped the TCE into above-ground storage tanks and later
into drums for resale to industrial users. The record shows that railcars delivering TCE
to Dico sometimes leaked TCE onto the ground. Tr. at 266-68. Furthermore, the
record reveals that some TCE leaked in the process of transferring the TCE from the
storage tanks to smaller drums for commercial distribution. Arthur Dale Hilliard, Jr.,
a former Dico employee, testified that he witnessed drums leaking TCE in the drum
storage area. Tr. at 263.




                                          -13-
                                 2. TCE Entered Soil

       Dico further suggests that even if TCE and TCE sludge leaked or was dumped
on its property, the District Court clearly erred in concluding that those TCE spills
entered the soil. Despite this assertion, the record contains evidence of at least three
sources of releases directly onto the soil: from railcars dripping TCE onto the ground,
from dumping TCE sludge on the ground and covering it with fill material, and from
spreading TCE as a dust suppressant directly onto the soil. 

pt. III.A.1.
While Dico asserts that Abdul’s testimony that any TCE spread or leaked on top of the
soil would only penetrate one foot deep is conclusive, other evidence contradicts his
assertion. Robertson testified that TCE is heavier than water and it sinks through the
soil unless impeded by some less-permeable barrier. He explained that TCE moves
more quickly than water through the soil because it is less viscous. Tr. at 38. Coupled
with Strouf’s testimony regarding the quantity of sludge buried and spread on Dico’s
property, the District Court did not clearly err in concluding that releases of TCE on
Dico’s property penetrated the soil to a depth greater than one foot.

        Dico also argues that the government offered no proof regarding how TCE
moved into the soil through the walls of the concrete containment pit holding the
degreaser or through the concrete slab in the drum storage area. Abdul testified that
TCE does not move through concrete at any significant rate; Robertson testified that
TCE penetrates concrete, both in its liquid and its vapor phases, because it is a heavy
chemical and is less viscous than water. Robertson noted that some of the highest
concentrations of contamination at the Site were found below these concrete slabs and
that, in his experience, high concentrations of TCE are often found under concrete-lined
containment pits. The expert testimony on this question of causation conflicted.
Weighing the testimony of both experts, the District Court found Robertson's testimony
more persuasive and credible. United States v. Dico, Inc., No. 4-95-10289, at 17-20
(S.D. Iowa Feb. 14, 2000) (Ruling on Liability and Order Setting Hearing Date).

                                          -14-
Nothing Dico has argued convinces us that the District Court clearly erred in so
concluding. See Lansford-Coaldale Joint Water Auth. v. Tonolli Corp., 

,
1218 n.7 (3d Cir. 1993) (concluding that the court would "not disturb the district court's
decision to credit the reasonable testimony of one of two competing experts,"
particularly where their differences rest on a different interpretation of the same data).
The record, taken as a whole, shows evidence in support of each basis for the District
Court's conclusion that Dico released TCE on its property. We hold that the court did
not clearly err in making this factual finding.

                                           B.

        Dico mounts a similar challenge to the court’s finding that TCE migrated through
the soil and into the groundwater on Dico’s property. Dico argues that the District
Court based its finding entirely on theories drawn from Robertson's testimony, and that
because his testimony was unreliable and should have been excluded, no basis exists
for the finding and it is clearly erroneous. As we previously concluded, the District
Court did not err in admitting Robertson's testimony. 

pt. II.A. Therefore,
Dico's challenge also fails. In response to Dico’s assertion that allowing the court to
rely on such "circumstantial evidence of proximity" ignores our opinion in Dico II, we
emphasize that our decision in that portion of this litigation merely reversed a grant of
summary judgment. We found a material issue of fact, appropriate for resolution at
trial, on the question of causation. Dico 

. We did not hold that,
once the government had been put to its proof on this claim, such evidence could never
form the basis of a judgment in favor of the EPA.

      In any event, our review of the record reveals that the government presented
evidence that TCE migrated through the soil into the groundwater, that Dico presented
evidence to the contrary, and that the District Court found the government's evidence
more compelling. Given the evidence the District Court had before it, we cannot say


                                           -15-
the court clearly erred in finding that Dico's releases of TCE migrated into the
groundwater and caused the contamination at issue.

                                          IV.
       Dico also challenges the District Court’s grant of summary judgment to the
government on the quantum of cleanup costs that would be awarded. The court
awarded the government $4,129,426.67 in costs attributable to the EPA's efforts to
remediate the Site's groundwater contamination. As part of these costs the court
awarded $508,284.76 for indirect costs,9 $730,060.74 for oversight costs,10 and
$370,453.57 for attorney fees and litigation costs. Dico contends that these costs are
not recoverable as a matter of law under CERCLA and the District Court should not
have allowed the government to recover any costs in these categories. We review this
question of statutory interpretation de novo. Braswell v. City of El Dorado, 

, 958 (8th Cir. 1999).

       CERCLA's cost recovery provisions specifically allow the government to bring
suit to recover "all costs of removal or remedial action incurred . . . [that are] not
inconsistent with the national contingency plan." 42 U.S.C. § 9607(a)(1)(A). The
statute defines "remedial action" as including

      such actions at the location of the release as storage, confinement,
      perimeter protection using dikes, trenches, or ditches, clay cover,
      neutralization, cleanup of released hazardous substances and associated
      contaminated materials, recycling or reuse, diversion, destruction,

      9
       Indirect costs include "costs expended for the operation of the Superfund
program to support site-specific cleanup activities." United States v. Dico, Inc., No.
4-95-10289, at 2 (S.D. Iowa Mar. 28, 2000) (order granting summary judgment on
response costs).
      10
       Oversight costs encompass "costs incurred by the EPA in overseeing activities
conducted by private parties other than EPA contractors." 

of reactive wastes, dredging or excavations, repair or
      replacement of leaking containers, collection of leachate and runoff,
      onsite treatment or incineration, provision of alternative water supplies,
      and any monitoring reasonably required to assure that such actions protect
      the public health and welfare and the environment.

(emphasis added). CERCLA further provides that "all such terms,"
including response, removal, and remedial action, "include enforcement activities
related thereto." 

                                            A.

        Dico contends that the definitions found in CERCLA do not encompass recovery
of the indirect and oversight costs incurred by the EPA in connection with the remedial
activities undertaken at the Site. Dico relies on the Third Circuit's opinion in United
States v. Rohm & Haas Co., 

(3d Cir. 1993), to argue that we must apply
the clear statement doctrine, as set forth in National Cable Television Ass'n v. United
States, 

(1974),11 to decide whether 42 U.S.C. § 9607(a) authorizes the


      11
         In National Cable, the Supreme Court reviewed fees that the Federal
Communications Commission sought to impose upon community antenna television
(CATV) systems. Pursuant to the Independent Offices Appropriation Act, the FCC
sought to impose fees on CATV systems that would reimburse the FCC for both the
direct and indirect costs of regulating the CATV industry. The Court struck down the
FCC's proposed fee structure. Imposing such regulatory costs on the CATV systems
would, in essence, allow the FCC to tax the systems. To avoid the constitutional
dilemma presented by such an interpretation, the Court defined the charges imposed
upon the systems as a "fee," which by definition must be an amount paid to the agency
for some benefit to the recipient "not shared by other members of society," such as
permission to conduct a particular 

. The fees could be based
only upon the FCC's direct costs, i.e., things that provided a direct benefit to individual
cable systems. But because the "fees" proposed by the FCC included reimbursement
for monies spent for the benefit of the general public, the Court remanded to the FCC
                                           -17-
recovery of oversight costs and indirect costs. In Rohm, the Third Circuit analyzed the
EPA's request pursuant to CERCLA for recovery of costs incurred in "overseeing a
hazardous waste cleanup performed and paid for by a private 

.
Both parties in Rohm agreed that the activities at issue fell under CERCLA's definition
of removal activities, rather than remedial activities. 

Therefore, the Rohm
court analyzed the recovery of oversight costs in the context of the 42 U.S.C. §
9601(23) definition of "removal." The court concluded that the statute lacks the clear
statement mandated by National Cable, and therefore oversight costs are not
recoverable by the EPA. 

.

       We respectfully disagree with the Third Circuit's analysis. The statutory scheme
at issue in National Cable involved the imposition of user fees on parties regulated by
the FCC for the benefit of conducting communications-related business. Conversely,
CERCLA is a remedial statute designed to make parties responsible for introducing
hazardous waste into the environment pay for cleaning up the messes they have
created. See United States v. Lowe, 

, 401-03 (5th Cir. 1997) (discussing
remedial nature of CERCLA). The provisions allowing the EPA to recover costs are
meant to make the guilty parties pay and thus are not like the user fees at issue in
National Cable. See 

(holding the National Cable analysis
to be inappropriate in the CERCLA context because CERCLA is a remedial statute;
it does not impose user charges on a regulated industry, and therefore the clear
statement doctrine does not apply to cost-restitution awards in CERCLA cases); Atl.
Richfield Co. v. Am. Airlines, Inc., 

, 568 (10th Cir. 1996) ("CERCLA
response costs are not user fees or taxes . . . . EPA oversight costs are not fees or taxes
levied against innocent members of a regulated industry to pay the EPA's general
administrative costs, but part of the damages caused or contributed to by specific
persons."); cf. United States v. Hyundai Merch. Marine Co., 

, 1189-90
(9th Cir.), cert. denied, 

(1999) (holding, in action brought pursuant to the


to revise the proposed fee structure.
                                           -18-
Oil Pollution Act of 1990, that language allowing the United States to recover "removal
costs" encompassed costs of monitoring Hyundai's private cleanup efforts; rejecting
Hyundai's argument that National Cable should apply by pointing out that recovery of
these types of costs does not constitute a tax). We are persuaded, as the Tenth Circuit
held in Atlantic Richfield, that "Rohm & Haas departed significantly from prior case
law that had construed the cost recovery provisions of CERCLA 

. We decline to adopt the Third Circuit's narrow approach. Accord United States
v. Chromalloy Am. Corp., 

, 349 (5th Cir. 1998) ("The government's
oversight costs in a responsible party clean-up are response costs under CERCLA.").

       Furthermore, even if we were to apply the National Cable clear statement
doctrine, we would conclude that its requirements are met with regard to CERCLA's
cost recovery provisions for remedial (as opposed to removal) actions. As the court
pointed out in Atlantic Richfield, the Third Circuit only addressed CERCLA's language
defining removal actions under § 9601(23). CERCLA defines remedial actions more
broadly to include "any monitoring reasonably required to assure that such actions
protect the public health and welfare and the environment." 42 U.S.C. § 9601(24)
(emphasis added). This language provides the specific congressional delegation of
authority to the EPA that the clear statement doctrine of National Cable, as interpreted
in Rohm & Haas, seems to require. Thus, even under the more restrictive approach
advocated by Dico, we would conclude that oversight and indirect costs are
recoverable in remedial actions under CERCLA.

                                          B.

       Dico also contends that the definitions found in CERCLA do not encompass
recovery of the government’s attorney fees, and that even if attorney fees are
authorized the government bears the burden of proving that the fees requested are
reasonable. We reject Dico's challenge on both grounds. First, the language of the
statute provides that attorney fees are recoverable as response costs under CERCLA.

                                          -19-
See 42 U.S.C. § 9601(25) (stating that the terms response, removal, and remedial
action "include enforcement activities related thereto"); United States v. Gurley, 

, 1199-1200 (8th Cir. 1994), cert. denied, 

(1995) (affirming
award of attorney fees to government under 42 U.S.C. § 9601(25) and § 9607(a)(4)(B)
for work done by EPA legal staff and Department of Justice attorneys).12

       Second, we reject Dico's argument that the government bears the burden of
proving that its requested attorney fees are reasonable. We specifically addressed this
argument in United States v. Northeastern Pharmaceutical & Chemical Co., 

(8th Cir. 1986) (NEPACCO), cert. denied, 

(1987), and held that the
party claiming that costs are unreasonable has the burden of proof. We examined the
statutory cost-recovery language applicable to the government's request for
reimbursement of costs, and noted that it provides that the "government may recover
from responsible parties 'all costs of removal or remedial action . . . not inconsistent
with the'" National Contingency Plan (NCP). 

(emphasis
added) (quoting 42 U.S.C. § 9607(a)(4)(A)). We concluded that this language created
the conclusive presumption that all costs incurred by the government that are not
inconsistent with the NCP are, in fact, reasonable costs. 

Dico therefore
bears the burden in this litigation of proving that the government's requested recovery
costs, whether attorney fees or otherwise, are inconsistent with the NCP.13 See also


      12
        We note that the Supreme Court's decision in Key Tronic Corp. v. United
States offers little guidance on the issue of the government's recovery of attorney fees
as part of its "enforcement activities" under CERCLA. 

, 819 (1994)
("Though we offer no comment on the extent to which [enforcement activity] forms the
basis for the Government's recovery of attorney's fees through § 107, the term
'enforcement activity' is not sufficiently explicit to embody a private action under § 107
to recover cleanup costs.").
      13
        Dico argues that the Ninth Circuit's decision in United States v. Chapman, 

(9th Cir. 1998), required the District Court to limit the government's attorney
fees award to those fees reasonably, not actually, incurred. We note that the decision
                                           -20-
Chromalloy Am. 

n.3 ("The burden of proving inconsistency
with the NCP rests with the responsible party."). Presumably, any attorney fees that
were not reasonably incurred would be held inconsistent with the NCP (else there
would appear to be no limiting principle to a claim by the government for attorney fees
as part of CERCLA response costs), but the burden would be on the responsible party
to show unreasonableness. Dico has not attempted to show that the fees requested are
inconsistent with the NCP.

      We affirm the District Court's grant of summary judgment awarding the
government a recovery of its costs incurred in connection with the cleanup of the Site
and establishing the amount of that recovery.14

                                          V.

       Before this case went to trial, Dico moved to add two affirmative defenses to its
answer. Dico sought to argue that the EPA's claim would "effect an unconstitutional
taking of private property" in violation of the Takings Clause of the Fifth Amendment,
and would "deprive Dico of property without due process of law" in violation of the



of the Ninth Circuit is not binding on this Court, and, though the decision reaches a
result that we find attractive, we must respectfully reject that court's analysis of the
CERCLA fee issue. The Ninth Circuit applied the Supreme Court's decision in Hensley
v. Eckerhart, 

(1983), to the award of fees under CERCLA. The facts of
Hensley are quite distinguishable—the fee award at issue was determined under the
prevailing-party provisions of 42 U.S.C. § 1988. In this Circuit, we have not applied
Hensley to an award of fees under CERCLA. CERCLA makes no mention of fee
awards to "prevailing parties." Moreover, the Act's language indicates that fee awards,
as with other response costs, must merely be consistent with the NCP.
      14
        Furthermore, we find Dico’s argument that a material issue of fact remains
regarding the calculation of attorney fees, based upon pending litigation against the
Department of Justice, to be completely without merit.
                                          -21-
Due Process Clause of the Fifth Amendment. In essence, Dico's proposed defenses
challenge the retroactive application of CERCLA to Dico's activities. The United
States moved under Federal Rule of Civil Procedure 12(f) to strike these defenses as
foreclosed by our decision in 

(upholding the
constitutionality of retroactive application of CERCLA under both the Due Process and
Takings Clauses of the Fifth Amendment). The District Court granted the United
States's motion. We review for an abuse of discretion. See Nationwide Ins. Co. v.
Cent. Mo. Elec. Coop., No. 00-2012, 

, at *5 (8th Cir. July 31, 2001);
see also Stanbury Law Firm v. IRS, 

, 1063 (8th Cir. 2000) (noting that
district courts enjoy "liberal discretion" under Rule 12(f)).

       Dico argues that the Supreme Court's decision in Eastern Enterprises v. Apfel,

(1998) (addressing Fifth Amendment challenge to Coal Industry Retiree
Health Benefit Act of 1992), supersedes our due process and takings holdings in
NEPACCO. We reject Dico's argument for two reasons. First, no single Fifth
Amendment rationale commanded a majority of the Court's votes in Eastern. Four
Justices voted to strike down the statute as unconstitutional based upon a Takings
Clause 

. The fifth vote for striking down the statute rejected
the takings analysis and held that the statute violated the Due Process Clause. 

(Kennedy, J., concurring in the judgment and dissenting in part). Second, the
background and purpose of the Coal Industry Retiree Health Benefit Act differs greatly
from that of CERCLA; Congress intended CERCLA to apply retroactively and acted
purposefully to allocate the cost of hazardous waste cleanups sites "to those who were
responsible for creating the sites." Franklin County Convention Facilities Auth. v. Am.
Premier Underwriters, Inc., 

, 552 (6th Cir. 2001) (holding that Eastern has
no precedential effect because no majority agreed on a rationale for striking down the
statute, and also holding that Eastern has no effect on the constitutionality of
retroactively applying CERCLA's provisions).

    In sum, we are unpersuaded by Dico's challenge to CERCLA on Fifth
Amendment grounds. We previously resolved the constitutionality of CERCLA's
                                         -22-
retroactive application in NEPACCO, and we are unpersuaded that the Supreme
Court's decision in Eastern undermines NEPACCO. The District Court did not abuse
its discretion in granting the government's motion to strike these constitutional defenses
from Dico's answer. Cf. United States v. Winnebago Tribe, 

, 1007 (8th
Cir. 1976) (upholding district court's striking of defense where federal statute clearly
authorized appellee's activities).

                                           VI.

       In summary, we reject Dico's challenges to the District Court's evidentiary
rulings and to its factual findings on causation regarding Dico's liability for TCE
contamination at the Site. The District Court appropriately granted summary judgment
on the government's request for its costs incurred in connection with the cleanup at
Dico's facility, and did not err in striking Dico's constitutional defenses. The judgment
of the District Court is affirmed.



      A true copy.

             Attest:

                 CLERK, U.S. COURT OF APPEALS, EIGHTH CIRCUIT.




                                           -23-