IN THE UNITED STATES COURT OF APPEALS

                            FOR THE FIFTH CIRCUIT



                                  No. 95-60228



MARINE SHALE PROCESSORS INC.,

                                                   Petitioner,

                                       versus

UNITED STATES ENVIRONMENTAL PROTECTION AGENCY,

                                                   Respondent.


                    Petition for Review of Order of
                     Environmental Protection Agency



                                 April 18, 1996

Before REYNALDO G. GARZA, KING, and HIGGINBOTHAM, Circuit Judges.

HIGGINBOTHAM, Circuit Judge:

     This case is an appeal of Marine Shale Processors, Inc. from

final   agency    action    of   the    Environmental      Protection     Agency.

Specifically,      MSP    challenges     EPA’s    decision    to   deny      MSP’s

application for a Boiler and Industrial Furnace Permit required by

the Resource Conservation and Recovery Act, 42 U.S.C. § 6901-92K.

This case is one of the trio described in United States v. Marine

Shale Processors, Inc., No. 94-30664.            We affirm.



                                         I

     In   1980,     EPA    promulgated       regulations   pursuant     to    RCRA

governing the treatment, storage, and disposal of hazardous waste.
See, e.g., Final Rule, Hazardous Waste Management:                       Overview and

Definitions; Generator Regulations; Transporter Regulations, 45

Fed. Reg. 12,721 (1980); Final Rule, Interim Final Rule, and

Request      for     Comments,   Hazardous           Waste      Management      System:

Identification and Listing of Hazardous Waste, 45 Fed. Reg. 33,082

(1980). These regulations defined two methods of processing waste,

incineration and recycling.           The rules required facilities engaged

in incineration to procure a permit called a Subpart O permit, a

reference to 40 C.F.R. pt. 264 subpt. O.                      See Proposed Rule and

Request for Comment, Identification and Listing of Hazardous Waste;

Amendments to Definition of Solid Wastes, 53 Fed. Reg. 519, 522

(1988).       Facilities engaged in recycling could operate without

permits.      See 45 Fed. Reg. at 33,120 (promulgating 40 C.F.R. §

261.6); see also Final Rule, Hazardous Waste Management System;

Definition of Solid Waste, 50 Fed. Reg. 614, 626-27 (1985).

       In    1985,   EPA   defined    a    new   category       of   hazardous    waste

processing devices called "industrial furnaces," a term defined to

include "aggregate kilns" having certain characteristics.                       50 Fed.

Reg.    at    661.      Industrial        furnaces      could    engage    in    either

incineration or burning for energy recovery.                      If the industrial

furnace facility engaged in incineration, then it needed a Subpart

O permit.         If the industrial furnace engaged in recycling, no

permit      was   necessary.     50    Fed.      Reg.    at    626-27.     MSP   began

operations in 1985, claiming an exemption from the Subpart O permit

requirement on the grounds that its kiln was an aggregate kiln and

that its facility was an industrial furnace engaged in recycling.


                                            2
     On August 14, 1990, the United States sued MSP in United

States District Court for the Eastern District of Louisiana in the

action giving rise to Nos. 94-30419 and 94-30664, claiming among

other things that MSP had incinerated hazardous waste without a

Subpart O permit since it opened for business in 1985.              In 1991,

EPA promulgated new rules requiring that all devices using thermal

combustion to treat hazardous wastes have either a Subpart O permit

or a new form of permit for recycling facilities called a Boiler

and Industrial Furnace permit.         Final Rule, Burning of Hazardous

Wastes in Boilers and Industrial Furnaces, 56 Fed. Reg. 7134, 7138

(1991).    These regulations ended the exception from the permit

requirement for facilities engaged in recycling.             MSP submitted a

BIF permit application and a Certification of Compliance with BIF

regulations. On the basis of these filings and its contention that

it fit within the previously existing recycling exemption, MSP

claimed interim status to operate while EPA considered the permit

application. EPA's internal consideration of MSP's application for

a BIF permit proceeded simultaneously with litigation of the United

States' action in Louisiana District Court.

     On January 31, 1994, EPA issued a tentative decision denying

MSP's BIF permit application.          EPA rested its tentative denial

decision upon its conclusion that MSP did not produce aggregate and

that its    system   did   not   use   thermal   treatment    to   accomplish

recovery of materials or energy within the meaning of 40 C.F.R. §

260.10.    EPA opened its decision for public comment.




                                       3
      A jury trial on the United States' claim in district court

that MSP had incinerated waste without a permit began in April,

1994.   At the end of a five-week trial, the court submitted 13

interrogatories to the jury.       In late May, the jury found itself

able to agree to the answer to only nine of the questions.               The

questions relevant to this appeal, together with the jury's answer

if any, were as follows:

           1. Was MSP entitled to a recycler exemption from
      the requirement of a permit as an operator of an
      incinerator of hazardous waste? (unable to answer)

           2. Were all of the hazardous wastes accepted by MSP
      beneficially used or reused or legitimately recycled?
      (unable to answer)

           2(a). Were all of the hazardous wastes accepted by
      MSP prior to August 21, 1991, beneficially used or reused
      or legitimately recycled? (unable to answer)

           10.   Is MSP's rotary kiln an aggregate kiln? (yes)

           13. Are the rotary kiln, oxidizers Nos. 1 and 2,
      and slag box part of a kiln system that produces
      aggregate? (yes)

Because the jury failed to answer four of the interrogatories, the

district court declared a mistrial.

      In September, 1994, EPA issued a final decision denying MSP's

application for a BIF permit.        EPA rested upon its finding that

MSP's rotary kiln system did "not meet the definition of aggregate

kiln and, therefore, does not meet the definition of industrial

furnace." EPA also cited MSP's poor history of compliance with the

environmental laws, as well as its finding that MSP could not

qualify as an aggregate kiln because it destroyed hazardous waste.

MSP   appealed   to   the   Environmental   Appeals   Board,   relying    on


                                     4
principles    of   Article    III,    the    seventh     amendment,    collateral

estoppel, due process, and the Administrative Procedures Act, 5

U.S.C. §§ 701-06.

     In March, 1995, after a review of the record, the EAB affirmed

EPA's denial.         In re Marine Shale Processors, Inc., Dkt. No.

06900009, RCRA Appeal No. 94-12, 

(EPA 1995).                   The

EAB stated that MSP did not produce "commercial-grade aggregate"

from its system and thus that its facility could not qualify as an

aggregate    kiln.      The   EAB    questioned    EPA's    reliance    on   MSP's

compliance history and on MSP's destruction of hazardous waste, but

ultimately affirmed the decision in its entirety.              In April, 1995,

EPA finally denied MSP's BIF permit application on all grounds

stated in its September, 1994 ruling.              MSP appeals the denial of

its permit, invoking our authority under 5 U.S.C. § 706(2) to set

aside final agency action.          We affirm.



                                        II

     MSP invokes Article III, the Seventh Amendment, and collateral

estoppel principles to attack EPA’s permit denial.

                                        A

     MSP     begins    its    assault       upon   the    permit   denial    with

constitutional arguments based on Article III and the Seventh

Amendment.    Its first argument is that Article III and the Seventh

Amendment prevent EPA from ruling on its permit application.                  Its

second argument is that the United States, by filing its lawsuit

and thus invoking the judicial power of an Article III court, could


                                        5
not continue to consider in an internal administrative proceeding

issues identical to those being litigated in the Article III court.

With cites to Montesquieu and Madison, MSP argues that the moment

the United States filed suit the district court obtained exclusive

power to decide any issue before it and that EPA’s permitting staff

could not resolve any legal question pending before the district

court without running afoul of the constitutional prohibition

forbidding Executive Branch review of Article III court decisions.

In a similar vein, MSP invokes the Seventh Amendment, contending

that once the Seventh Amendment is activated as to an issue, a

party is entitled to have the issue resolved by a jury.

      With regard to both MSP’s Seventh Amendment and Article III

arguments, we begin with the proposition that, in the absence of a

simultaneous district court proceeding, Congress violated neither

constitutional principle by providing that EPA should adjudicate

MSP’s permit application. See In re Texas General Petroleum Corp.,

, 1336 (5th Cir. 1995) (“Whether an Article III court

is necessary involves the same inquiry as whether a litigant has a

Seventh Amendment right to a jury trial.”) (citing Granfinanciera,

S.A. v. Nordberg, 

, 53-54 (1989)).        MSP’s contention to

the   contrary   comes   decades,       perhaps   centuries,   too   late.

Congress’s choice to grant EPA authority over the permit proceeding

represents a classic constitutional example of the public rights

doctrine.

      Viewing our inquiry as governed by “practical attention to

substance rather than doctrinaire reliance on formal categories,”


                                    6
Thomas v. Union Carbide Agricultural Products Co., 

,

587 (1985), we note the following characteristics of the permit

proceeding.    First, it is a dispute to which the sovereign is a

party.     See Crowell v. Benson, 

, 50 (1932); Murray’s

Lessee v. Hoboken Land & Improvement Co., 59 U.S. (18 How.) 272,

284 (1855).      Second, the permit requirement “serves a public

purpose as an integral part of a program safeguarding the public

health.”     

.         Third, the scientific and

technical nature of the decisions in this permit proceeding make

the decision “peculiarly suited to examination and determination by

an   administrative    agency    specially    assigned     to   that     task.”

. Fourth, the permit proceeding, and indeed

most of RCRA itself, deals with a narrowly cabined area of the law.

See Commodity Futures Trading Commission v. Schor, 

,

852 (1986) (referring to a “particularized area of law”) (internal

quotation marks omitted). Fifth, the permit application implicates

a federal right not immediately analogous to the state common law

causes of action “assumed to be at the ‘core’ of matters normally

reserved to Article III 

.         Sixth, EPA,

“unlike the bankruptcy courts under the 1978 Act, does not exercise

‘all ordinary powers of district courts,’ and thus may not, for

instance,    preside   over   jury   trials   or   issue   writs    of   habeas

(quoting Northern Pipeline Construction

Co. v. Marathon Pipe Line Co., 

, 85 (1982) (plurality

opinion)). Seventh, Congress has provided for judicial review over

the agency’s permit denial under 5 U.S.C. §§ 704-06.            See 

at 854-55. Under such circumstances, Congress violated neither the

Seventh Amendment nor Article III in delegating the permitting

decision to EPA.

     Having established that EPA could constitutionally adjudicate

the permitting proceeding in the absence of a suit by the United

States, we address MSP’s argument that the enforcement action

altered the constitutional landscape.         We find MSP’s contention

convincing in neither the Article III nor the Seventh Amendment

context. At bottom, both arguments fail for the same reason:          This

is not a case in which EPA has sought to review or alter the

decision in the district court, to reverse the district court’s

findings, or to interfere with the judiciary’s ability to issue a

binding decision.    Cf. Hayburn’s Case, 2 U.S. (2 Dall.) 40 (1792)

(suggesting that the judiciary could not render an opinion as to

whether a citizen was entitled to a pension when both the Secretary

of War and the Congress retained power to decide whether to honor

the judiciary’s judgment); Town of Deerfield v. FCC, 

(2d Cir. 1993) (rejecting the FCC’s attempt to alter or amend a

federal court judgment).       EPA has, to be sure, disagreed with

several of the conclusions of the district court, but it has never

sought to interfere with the effect that these conclusions have

upon the causes of action being adjudicated there.            The permit

application   and   the   district   court   litigation   involve   common

issues, but the two proceedings are deciding different questions,

the most important of which is that the permit proceeding concerns

whether MSP may operate legally in the future, while the district


                                     8
court proceeding concerns whether MSP has operated legally in the

past.

     Should EPA attempt to use the rulings in the permit proceeding

to collaterally estop MSP in the district court action, MSP’s

argument may have force; we make no comment on this question.         Cf.

Parklane Hosiery Co v. Shore, 

(1979) (holding that the

Seventh Amendment does not prevent an equity court’s determination

of legal issues from estopping relitigation of the same issues in

a   subsequent   action   at   law).   But   given   that   EPA     could

constitutionally adjudicate the permit application in the absence

of the district court litigation, MSP can make no argument until

the permitting proceeding has some effect upon the issues being

decided in the district court litigation.      When and if MSP feels

that effect, it may litigate these issues.

     Accordingly, EPA has never sought to judge its own case any

more than it does in any normal permit application proceeding.        If

the permitting arm of the agency could constitutionally exercise

jurisdiction over MSP’s permit application in the absence of a

district court proceeding, as we have held, then nothing in Article

III prevented EPA’s permit staff and the district court from

proceeding   simultaneously.      Having   established   this     general

principle, we examine MSP’s arguments in detail.

                                   1

     MSP argues that as a result of EPA’s suit, Article III gave

the district court exclusive power to decide the controversy

between MSP and EPA.      MSP points to no statute or constitutional


                                   9
provision granting exclusive jurisdiction to the federal district

courts to decide all disputes between EPA and entities like MSP.

Its argument assumes that there would be no Article III bar had EPA

denied the permit and then filed the district court suit, or if EPA

had litigated the suit to completion and then denied the permit.

     We find MSP’s Article III arguments unconvincing.        State

courts are not Article III courts, yet nothing in Article III

prevents a state court from litigating the same controversy pending

before a district court. Kline v. Burke Construction Co., 

, 230 (1922).   In such cases, if the state court reaches final

judgment first, its disposition may preclude further litigation in

the district court without violating Article III.    

Wayside Transportation Co. v. Marcell’s Motor Express, Inc., 

, 870-71 (1st Cir. 1960).    MSP cites to no case suggesting

that this principle would change if the state itself were a party

to both the state court and federal court litigation.      Like the

Sixth Circuit,

     We see no reason why the rule permitting a second
     tribunal to proceed to the decision of an in personam
     matter within its jurisdiction, in spite of the fact that
     another action between the same parties is pending in
     another tribunal, should not be applicable as between a
     United States District Court and a federal administrative
     agency   in   which   Congress   expressly   has   placed
     responsibility for regulation in a specific area.

Ashland Oil & Refining Co. v. FPC, 

, 21 (6th Cir. 1970).

     MSP cites California v. FPC, 

(1962), for the

broad proposition that any time a federal court has jurisdiction

over a controversy in which an agency is a party, the agency must

withhold administrative action until the court has reached a

                                10
decision.    We do not read California so broadly.               In that case, the

Supreme Court held that the FPC should not have approved a merger

while a suit challenging the merger, filed by the United States,

was pending in federal district court.                Justice Brennan’s majority

opinion does not mention Article III.                  Instead, it justifies its

holdings with “practical reasons,” primarily the difficulty of

unscrambling an already consummated merger.

      MSP points out no analogous practical considerations in this

case.      To the contrary, EPA could reasonably decide that the

district court litigation promised to continue for years. The EAB,

for instance, considered MSP official George Eldredge’s statement

that “whatever action EPA proposes, and whatever the outcome of the

lawsuit, the case is going to drag on for years.                     In the mean[

]time, we’ll be doing business as usual.”                     In re Marine Shale

Processors, Inc., Dkt. No. 0690009, RCRA Appeal No. 94-12, 

, at * 17 (EPA 1994) (internal quotation marks omitted). EPA

could conclude that awaiting the decision of the judiciary on those

issues common to the district court litigation and the permit

proceeding would unduly delay resolution of the important questions

in   the   latter    and   would      allow      an   admitted   violator    of   the

environmental       laws   to   continue        operating,    perhaps   in   further

violation of these laws, until the conclusion of the litigation and

the inevitable appeal.            Normally, speedy adjudication from an

administrative agency is something to be encouraged, and we cannot

fault   EPA’s   decision        not   to   await      the   unavoidably   ponderous

progress of the district court litigation.


                                           11
                                         2

     For    similar    reasons,     we    reject      MSP’s    seventh      amendment

argument.    As EPA acknowledges, MSP has a right to a jury trial in

the district court proceeding.           See Tull v. United States, 

(1987).   But because the permit application triggered a public

rights   dispute,     MSP    has   no    right   to    a    jury    trial    in   this

proceeding.    Atlas Roofing Co. v. OSHRC, 

, 455 (1977);

see also 

(noting that jury trials may be incompatible

with agency processes).            MSP cites no case holding that the

pendency of an action in one tribunal in which a jury trial right

attaches prevents another tribunal from proceeding without a jury.

Unless and until MSP is prevented from litigating its defenses in

the district court to a jury, no arguable jury trial violation has

occurred.

     Nothing in Beacon Theaters, Inc. v. Westover, 

(1959), requires a different result.                  In Beacon Theaters, the

plaintiff filed an action seeking an injunction prohibiting the

defendant from prosecuting an antitrust suit.                       The defendant

counterclaimed in a complaint stating the antitrust cause of action

and demanded a jury trial.          The trial judge scheduled the hearing

on the plaintiff’s request for injunctive relief ahead of the jury

trial on the defendant’s antitrust claim.                  The Supreme Court held

that the trial court abused its discretion in scheduling the equity

suit first    because       such   scheduling      would     have   the   effect    of

depriving the defendant of its right to a jury trial in the

counterclaim.    Crucial to this holding was the fact that modern


                                         12
rules of civil procedure allowed joinder and joint resolution of

multiple claims of multiple parties, thus in effect giving the

plaintiff an adequate remedy at law by joining all involved.                      See

509, 511; see also Dairy Queen, Inc. v. Wood, 

, 472-73 (1962) (interpreting Beacon Theaters in this manner);

Lytle v. Household Manufacturing, Inc., 

, 550-52 (1990)

(discussing the importance of the fact that legal and equitable

claims were tried in the same lawsuit in applying the Beacon

Theaters rule).      In this case, in contrast, the district court may

not grant a permit, and the adjudicatory arm of EPA may not grant

the relief sought in EPA’s district court complaint.                       A single

proceeding could not resolve both issues.

                                          B

     MSP    next    argues   that       EPA’s   permit     denial    violated     the

principle of collateral estoppel.               In order to invoke collateral

estoppel,   however,       “the   issue       under   consideration       [must   be]

identical to that litigated in the prior action.”                     Copeland v.

Merrill Lynch & Co., 

, 1422 (5th Cir. 1995).                      MSP’s

collateral estoppel argument fails because the jury was either not

asked to resolve or unable to resolve questions crucial to EPA’s

decision to deny MSP’s permit application.

     MSP    based    its   BIF    permit       application    to    EPA   upon    the

contention that its kiln system constituted an industrial furnace.

40 C.F.R. § 260.10 defines industrial furnace as “any of the

following    enclosed      devices      that    are     integral    components     of

manufacturing       processes     and    that     use    thermal    treatment      to


                                          13
accomplish recovery of materials or energy.”              The definition then

lists a series of twelve devices by name, which Judge Duplantier

referred to as the “twelve apostles.” “Aggregate kilns” are one of

the twelve apostles. As applied to this case, then, section 260.10

requires EPA to grant MSP a permit if Marine Shale’s kiln is (1) an

aggregate    kiln,   (2)        that    is    an   integral   component     of   a

manufacturing process, and (3) that uses thermal treatment to

accomplish recovery of materials or energy.

     In addition, paragraph thirteen of the definition provides

that EPA may add other devices to the list of the twelve apostles

after   consideration      of    five    enumerated    factors   with   a   sixth

catchall statement that EPA might consider “other factors, as

appropriate.”   The first of these five factors is “the design and

use of the device primarily to accomplish recovery of material

products.”   The remaining four generally focus on the inquiry of

whether a facility makes a product.

     In an interpretive step that MSP has not challenged, EPA

analyzed MSP’s permit application in terms of the five factors

articulated in paragraph thirteen and did not directly focus on the

narrative criteria preceding the list of the twelve apostles.                    We

note that the first of the five factors mimics the second of the

narrative criteria and that both focus on whether a facility

recovers energy or materials.

     EPA denied MSP’s permit application on the basis of its

finding that MSP’s kiln system met none of the five factors




                                         14
enumerated in paragraph thirteen.1            In particular, EPA discussed

extensively its grounds for finding that the majority of the

hazardous    waste    processed    by   MSP   contributed       nothing   to   the

production of a product and were not recovered or recycled, and

therefore that MSP did not use thermal treatment to recover energy

or materials.        In addition, EPA found that MSP’s kiln did not

produce aggregate after defining aggregate according to commercial

specifications.

     The jury’s findings covered only a portion of the industrial

furnace   inquiry     considered   in    terms    of   either    the   narrative

criteria and the twelve apostles or the five factors in paragraph

thirteen.    The jury found that MSP’s rotary kiln was an aggregate

kiln, and that MSP’s kiln, oxidizers, and slag box were part of a

system that produced aggregate.          For whatever reason, the jury was

not asked whether MSP uses thermal treatment to recover energy or

materials.     The interrogatories most analogous to the thermal

treatment inquiry were questions 2 and 2(a), which asked the jury

whether the hazardous wastes received by MSP were beneficially used

or reused or legitimately recycled.              The jury failed to reach a

     1
        Although the EAB affirmed Region VI’s initial decision to
deny the permit primarily upon the ground that MSP’s kiln was not
an aggregate kiln, it “recognize[d] that the Region based its
decision on other grounds as well” and clarified that “to the
extent we have not ruled on those other grounds, nothing in this
decision should be construed as preventing the Region from basing
its final permit decision on these other grounds.” 

,
at * 24. The final agency action of which MSP complains is Region
VI’s formal denial of its permit application. This denial makes
clear that one of the bases of Region VI’s denial was MSP’s failure
to meet the criteria stated above. MSP has not argued to this
court that our review is limited to the grounds articulated in the
EAB’s decision.

                                        15
verdict on these interrogatories.     Accordingly, EPA decided issues

that the jury did not, and collateral estoppel does not apply.

     MSP seeks to avoid the force of this argument by contending

that the jury decided whether MSP used thermal processes to recover

energy or materials when it decided that MSP produced aggregate.

This argument is structurally identical to SWP’s contention, which

we rejected in United States v. Marine Shale Processors, Inc., No.

94-30419, at 11, that producing a product is necessarily recycling,

and we disagree for the same reason here.      A reasonable trier of

fact could find that, to the extent that MSP produced aggregate, it

did so without recovering the energy or materials in the hazardous

wastes that it accepted.

     MSP’s brief suffers from the assumption that the only issue in

the permit application proceeding was whether its rotary kiln

constituted an aggregate kiln.   That assumption is incorrect.   Not

all aggregate kilns are industrial furnaces, as the narrative

criteria of the definition of industrial furnace and the first of

the five factors in paragraph thirteen make clear.2

    2
       We also reject MSP’s implication that the district court’s
ruling that MSP had interim status as a BIF estopped EPA from
denying the permit. Interim status is designed to last only until
EPA rules on a permit application.
     Because of our disposition of MSP’s collateral estoppel
argument on the grounds of lack of identity of the legal issues
involved, we do not reach EPA’s argument that the jury’s findings
cannot support collateral estoppel because they have not been
incorporated into a final judgment. We note, however, the tension
between the dictum in Recoveredge L.P. v. Pentecost, 

,
1295 (5th Cir. 1995), which suggests that jury findings are
sufficient to support collateral estoppel even if the verdict has
not yet resulted in a final judgment, and the holding of Avondale
Shipyards, Inc. v. Insured Lloyd’s, 

, 1272 (5th Cir.
1986), which suggests that partial summary judgments are

                                 16
                                   III

      MSP argues that EPA's findings of fact and conclusions of law

were arbitrary and capricious.      We do not agree.

      We limit our review in this case to sections III.A and III.B

of EPA’s September 15, 1994 Statement of Basis for Denial of Permit

Application by Marine Shale Processors, Inc.           The findings of fact

and   conclusions   of   law   included   in   these    two   sections   are

sufficient to uphold EPA’s decision.           Although we find none of

EPA’s findings of fact or conclusions of law in these two sections

arbitrary, capricious, or contrary to law, we focus our discussion

here on the evidence underpinning the finding that MSP’s system

does not use thermal processes to accomplish recovery of energy or

materials and on certain determinations EPA made in deciding that

MSP’s material does not qualify as aggregate within the meaning of

40 C.F.R. § 260.10.

                                    A

      EPA’s finding that MSP has not designed or used its facility

to accomplish recovery of material products and thus that MSP does



insufficient. Under Fed. R. Civ. P. 54(b), both jury findings and
partial summary judgments are subject to revision before entry of
final judgment. Even when, as here, a trial judge has rejected a
motion for a judgment as a matter of law seeking to upset the
jury’s findings, Rule 54(b) provides that this rejection is itself
subject to revision at any time before final judgment. We thus
have some difficulty justifying a rule, if in fact it exists in
this circuit, allowing issue preclusion to attach to jury verdicts
but not to summary judgments, when neither has been incorporated
into a final judgment. We do not seek to resolve this tension in
this case.
     Given our disposition of this case, we also do not reach EPA’s
argument that the different burdens of proof and nature of the
proceedings precludes invocation of collateral estoppel.

                                    17
not use thermal treatment to accomplish recovery of materials or

energy is not arbitrary or capricious. As our discussion will make

clear, EPA’s decisions are highly technical and scientific and are

not readily susceptible to lay review. Most of these decisions are

factual.     We   bear   these   considerations        firmly   in   mind   when

considering MSP’s request that we upset EPA’s conclusions in an

area in which Congress has chosen to trust the experts.

       Throughout this section, we assume that MSP produces something

its calls aggregate and that its kiln system is an integral

component of the process for this aggregate’s production. We focus

entirely on the question of whether EPA could conclude that MSP

does not use thermal treatment to accomplish recovery of materials

or energy, or that the design and use of MSP’s device is not

primarily to accomplish recovery of material products.

                                        1

       EPA considered evidence that MSP processed quantities of “lab

packs” containing wastes that could contribute nothing to the

production of a product.         The lab packs were packages of kitty

litter and other absorbent material surrounding glass or plastic

containers of toxic chemicals.          For example, Dr. Douglas Kendall,

an EPA chemist, used MSP’s manifests and Material Characterization

Data   Sheets   to   determine   that       MSP   processed   sulfur,   toluene

solution,   ammonium     hydroxide,     hydrochloric     acid   solutions    and

mixtures, nitric acid, and sulfuric acid.              Dr. Kendall confirmed

that these wastes do not release significant energy when burned

and, because they react to form gases at high temperatures, could


                                      18
not provide    bulk     for    MSP’s   product.     EPA    considered   similar

evidence regarding such materials as poisons, pesticides, other

acids, and bases; specific substances included methylene chloride,

trichlorotrifluoroethane,              chloroform,        perchloroethylene,

trichloroethylene,            nitric     oxide,      fluorotrichloromethane,

pentachlorophenol,            ethylenediamine,       formaldehyde,       carbon

tetrachloride, and phosgene.           MSP’s experts could not specify how

many of these substances contributed to a manufacturing process.

MSP’s   handling   of    these     substances     also    suggests   that   they

contribute nothing to production. MSP employee Annika Keslick told

EPA that MSP normally opened ten percent of these lab packs, and

MSP’s examination upon opening was limited to matching the name on

the glass or plastic container within the pack to the information

contained on the MCDS or manifest.

       EPA could find that MSP was not accomplishing recovery of

energy or materials from these wastes.               The composition of the

wastes themselves did not allow their combustion to contribute to

any production process, and we cannot understand how MSP could have

recovered energy or materials from these wastes without sampling

them to verify their contents. MSP’s only defense of its treatment

of the lab pack wastes is that the kitty litter and other packaging

provide mass for its aggregate.              EPA could conclude that this

argument misconstrues the regulations and is wrong as a matter of

law.    One does not recycle hazardous waste by placing that waste

into a container and then recycling the container.




                                        19
     MSP correctly points out that the lab packs constitute only

around one percent of the total volume of wastes processed at its

facility.   Nevertheless, the amount of material is significant in

absolute terms in that MSP process an average to three to four

hundred of the lab packs per week, and MSP’s cavalier treatment of

these “feedstocks” gives us pause when we consider the remainder of

MSP’s claim that all of its wastes contribute in some way to its

process.

                                   2

     EPA considered evidence that MSP processed large quantities of

waste with metal contaminants that contribute in no legitimate way

to any manufacturing process and thus that MSP’s use of these

wastes did not constitute recovery of energy or materials.              The

metallic composition of these wastes spanned the periodic table and

included highly variable quantities of lead, barium, cadmium, iron,

silicon, aluminum, manganese, copper, zinc, bromine, strontium,

calcium, and chromium.    MSP’s kiln did not destroy these metals.

The residue from the process of metal-bearing waste, which MSP

calls primary aggregate, normally required slagging to reduce

leaching potential.

     MSP suggests that it used these metals for two purposes.

First, all provide mass for the ultimate product.             Second, some

compounds from these metals had other properties useful to the

manufacturing process or the ultimate product.         Dr. Paul Queneau,

a metallurgical engineer, told EPA that iron oxide and alumina and

titanium are   “chain   formers,   and   they   very   much   enhance   the


                                   20
environmental stability of the slag.”              Other metallic oxides lower

the melting point of the mixture and decrease its “melt viscosity.”

       EPA’s disbelief of these justifications was not arbitrary or

capricious.    EPA scientists stated that the metal content of the

waste necessitated slagging before the ultimate product could be

legally    placed   on   the     ground    and    that     the    slagging    process

significantly reduced the mass produced.                 Dr. Terrance McNulty, an

expert in extractive metallurgy, also provided evidence that many

of these metals impeded production of the slag.                         Barium, for

instance, which at times constituted fourteen to sixteen percent of

the slag mass, impeded production because the high melting points

of   its   compounds     made     liquification          more    difficult.      Most

importantly, EPA considered evidence suggesting that while many of

the metal compounds do exhibit some of the desirable properties

that Dr. Queneau identified, they do so only when present in

certain concentrations.         Chemist Stanley Wrobleski confirmed that

Marine Shale made no attempt to control the metallic composition of

its primary or slagged material and that metal concentrations

varied widely.      Moreover, EPA considered evidence such as a letter

from    Woodward-Clyde     Consultants,          MSP’s     primary    environmental

consultant, to George Eldredge, an MSP officer, stating that many

of   the   metal    compounds     “are    not     introduced       specifically     or

purposefully   into      the    raw   product     in     order   to   incorporate    a

particular physical characteristic into the produced aggregate but

are inherent elements of the raw materials used in the manufacture




                                          21
of the aggregate.”3   Under such circumstances, EPA could conclude

that MSP’s process did not recover of these metal-bearing wastes or

their metallic constituents.4

                                 3

     The largest percentage of MSP’s wastes consists of soil

contaminated by organic compounds.    MSP contends that the soil

        3
         Although this same letter concluded that “these elements
are beneficial in enhancing the quality of the produced aggregate,”
it appears that this conclusion was based entirely upon the
coincidence that “the majority of the elements of the produced
aggregate are also the major constituents in some of the more
common and select construction materials in use today.” Nothing in
this letter sought to match the concentration of metals in MSP’s
material to that in the more common and select construction
materials. Marine Shale’s argument would lead to the conclusion
that any material containing sugar, butter, eggs, and flour is a
cake.
    4
       EPA warned the regulated community about this type of “use”
of metallic compounds shortly after filing the lawsuit in this
case.

     The Agency notes in addition that it ordinarily does not
     consider metal-bearing hazardous wastes to be used as
     ingredients when they are placed in industrial furnaces
     purportedly to contribute to producing a product. . . .
     To be considered legitimate use as an ingredient, it
     would normally need to be demonstrated to EPA (or an
     authorized State) pursuant to § 261.2(f) that the
     hazardous metal constituents in the waste are necessary
     for the product (i.e. are contributing to product
     quality) and are not present in amounts in excess of
     those necessary to contribute to product quality. This
     would normally require some demonstration that these
     hazardous metal constituents do not render the product
     unsafe for its intended use. (The other sham recycling
     criteria discussed frequently by EPA would have also to
     be satisfied). The types of uses of hazardous wastes in
     industrial furnaces to produce waste-derived products of
     which the Agency is aware, such as using hazardous wastes
     to produce aggregate or cement[,] . . . do not appear to
     satisfy these criteria.”

Final Rule, Burning of Hazardous Wastes in Boilers and Industrial
Furnaces, 56 Fed. Reg. 7133, 7185 (1991).

                                22
provides raw material, or mass, for its aggregate and that the

organic compounds release heat when burned.        Accordingly, MSP

argues that both the soil and the waste contribute to its aggregate

production process.

     EPA’s rejection of these arguments was not arbitrary or

capricious.   EPA considered evidence that some of these wastes

consisted of soil contaminated with pentachlorophenol, which it

specifically labeled a low energy hazardous waste constituent.   In

addition, EPA could conclude that MSP’s process generates heat far

in excess of that needed to make its product.     Ronald Corwin, an

EPA expert witness, suggested that the vast majority of the heat

MSP produced from its burning travels in non-contact cooling water

through MSP’s facility and out into Bayou Boeuf.          While MSP

correctly points out that no recycling process is one hundred

percent efficient, EPA’s assessment of whether this heat is used or

wasted is a particularly technical judgment about the overall

efficiency of MSP’s process.    We will not disturb this judgment in

this case.

                                  4

     At oral argument, MSP strenuously contended that EPA’s permit

denial decision rested on the conclusion that EPA could reject the

application if MSP burned a thimbleful of hazardous waste for

destruction, and thus that EPA had imposed an unreasonable burden

in requiring a potential BIF to prove that it was recovering every

atom or every bit of heat from waste in order to claim entitlement

to a BIF permit.      We make no comment on this argument; this is


                                 23
simply   not   a   thimbleful   case.     EPA   has   concluded   that   the

overwhelming majority of MSP’s wastes are burned for destruction,

not used for recovery of energy or materials.         The findings of fact

and conclusions of law underlying these decisions are not arbitrary

or capricious.      EPA could conclude that to the extent that MSP

produced a product, it did so in spite of the wastes it purported

to recycle.

                                     5

     MSP’s final attack on this analysis is that a focus upon

recovery of energy or materials constitutes an analysis of the role

that each material plays in the manufacturing process and of the

purpose the particular facility serves.         After the promulgation of

the BIF regulations, MSP argues, a focus on purpose is improper.

In particular, MSP quotes the EAB’s statement that “we have serious

doubts as to whether after promulgation of the BIF rule the purpose

for which MSP is burning hazardous waste at the facility is

relevant to the determinant of whether MSP’s facility meets the

industrial furnace definition.”          In re Marine Shale Processors,

Inc., Dkt. No. 06900009, RCRA Appeal No. 94-12, 

, at

* 25 n.32 (EPA 1995).      MSP also notes that 40 C.F.R. § 266.100

establishes that the BIF rules regulate BIFs without regard to

whether the particular facility is burning for destruction or is

recycling.

     MSP’s argument fails to separate two analytically distinct

issues and regulations.         40 C.F.R. § 260.10 governs whether a

facility definitionally qualifies as a BIF.           Once a facility has


                                    24
definitionally qualified as a BIF, 40 C.F.R. pt. 266 subpt. H

governs most aspects of its operations, including burning for

destruction. Although we note that some tension might arise if EPA

were to interpret section 260.10's definition of BIF to exclude a

facility that burns a thimbleful of waste for destruction, EPA has

not done so here, as is made clear by EPA’s focus on whether MSP

uses its kiln system “primarily to accomplish recovery of material

products.”      40 C.F.R. § 260.10 (emphasis added).               We cannot

conceive   of   an   interpretation    of    “to    accomplish   recovery   of

materials or energy” and other similar phrases in section 260.10

that does not focus on purpose.

     An analysis of the preambles to the regulations defining BIFs

supports our conclusion.      In distinguishing between boilers and

incinerators, EPA did seek to shift the initial focus of the

definitional inquiry from primary purpose to structural design.

Thus, EPA considered and eventually adopted a definition of boiler

depending on whether the facility “achieve[s] heat transfer within

the combustion chamber itself, generally by exposing the heat

recovery surface to the flame.”            Proposed Rule, Hazardous Waste

Management System:      General, 48 Fed. Reg. 14,472, 14,483 (1983);

see Final Rule, Hazardous Waste Management System; Definition of

Solid Waste, 50 Fed. Reg. 614, 626-27 (1985).5           But EPA recognized

that some facilities normally engaging in recycling lacked this

     5
        Even in the boiler context, EPA            used the integral design
test as a proxy for discovering the                 primary purpose of the
facility.   See, e.g., 50 Fed. Reg. at              626 (“The definition of
boilers focuses on physical indicia of             their legitimate use for
energy recovery.”).

                                      25
distinguishing characteristic of boilers, and therefore chose to

rely in part upon the primary purpose test in defining industrial

furnaces.    50 Fed. Reg. at 626-27.       Thus, the language of the rules

and the preambles support our conclusion that EPA may interpret 40

C.F.R. § 260.10 to include a focus on the primary purpose of the

facility or the role played by wastes processed within it.

                                     B

     We hold that EPA’s refusal to label MSP’s kiln an aggregate

kiln was not arbitrary or capricious.            MSP’s primary attack upon

this portion of EPA’s reasoning is that EPA erred by narrowing its

definition of “aggregate” to “commercial grade aggregate.”                In

particular, MSP disputes EPA’s reliance upon standards promulgated

by the Louisiana Department of Transportation in reaching its

decision that MSP’s material does not constitute commercial grade

aggregate.        EPA’s interpretations of its own regulations are

entitled to substantial deference.               Ford Motor Credit Co. v.

Milhollin, 

, 566 (1980).          We find no error.

     In making its adjudicative decision, EPA had to employ some

set of standards to distinguish aggregate from any material, like

cigarette ash, capable of occupying space.            The history of EPA’s

focus on recycling of hazardous wastes to produce a commercial

product,     as    well   as   the   use    of     commercial   terms   like

“manufacturing” process and “industrial” furnace, suggests that

EPA’s decision to employ commercial criteria in its decision was

reasonable at least.




                                     26
     EPA considered factual evidence from witnesses knowledgeable

in the construction field that aggregate purchasers typically

employ at least some specifications for the product they purchase.

In addition, EPA heard evidence that a material must meet LaDOT

specifications before the State of Louisiana will buy it for state

construction projects and that many private commercial contractors

adopt these specifications as well.     In the face of this legal

history, regulatory language, and factual evidence, we cannot fault

EPA’s choice to rely on common commercial specifications to define

the term “aggregate kiln” in 40 C.F.R. § 261.10.

     We also find nothing wrong with EPA’s decision to consider

LaDOT criteria as highly indicative, though not conclusive, of the

content of common commercial specifications.   MSP has proposed no

alterative set of standards or definition.     Federal courts give

deference to an agency’s interpretation of its own rules; under the

circumstances of this case, however, we seriously doubt that such

deference is necessary.   We find no error in EPA’s choice of legal

standard.

     Given EPA’s legal interpretation of its own regulation, we

find nothing arbitrary or capricious in its application of this

interpretation to the facts at hand, and we refuse to upset its

conclusion that MSP’s material is not commercial grade aggregate.

MSP concedes that its material, and substances made from it, could

not comport with many of the LaDOT standards.      In addition, MSP

concedes that it conducts no tests at all on its material to

determine strength, size, shape, specific gravity, absorbency,


                                27
durability, compaction, or texture.           Although MSP presented expert

studies suggesting that its slagged and primary material could be

useful   in   the    production    of    certain   concrete     and   asphaltic

products, other experts disagreed.              The choice of which expert

opinions to credit belongs to the EPA permitting staff.                Like the

Environmental Appeals Board, we are struck by the fact that MSP has

never field tested any of the products that its experts testified

might possibly be manufactured in part from its slagged and primary

material and that none of MSP’s product has ever been commercially

used for these purposes. In re Marine Shall Processors, Inc., Dkt.

No. 06900009, RCRA Appeal No. 94-12, 

, at * 12 (EPA

1994).   Under such circumstances, EPA’s application of the law to

the facts is not arbitrary or capricious.



                                        IV

     MSP continues its attack on the permit denial process by

alleging   that     EPA's    failure    to   insulate   fully   the   personnel

litigating the district court action from those participating in

the permit denial process rendered the latter deficient under the

Due Process Clause.         MSP highlights the roles of two individuals,

Dr. Allyn Davis and Ms. Terry Sykes.6              We find no due process

violation.

                                         A


    6
       MSP also devotes a footnote to an allegation of misconduct
by Mr. Steven Silverman, an attorney in EPA’s Office of General
Counsel, labeling him the “finalizer” of EPA’s permit denial
decision. This portion of MSP’s argument lacks merit.

                                        28
      Dr. Davis was the Director of EPA Region VI's Hazardous Waste

Management Division, which has oversight of both enforcement and

permitting issues within Region VI.                  Dr. Davis referred MSP’s

facility to EPA's enforcement arm.                  He later made the initial

determination that MSP should not receive a BIF permit.                        MSP also

attacks Dr. Davis's adjudicative role on the grounds that his

deposition testimony showed that he had prejudged certain key

issues.

      We   find    nothing    remarkable       in   Dr.    Davis’s      role    in   the

permitting process.

      It is also very typical for the members of administrative
      agencies to receive the results of investigations, to
      approve the filing of charges or formal complaints
      instituting enforcement proceedings, and then to
      participate in the ensuing hearings.        This mode of
      procedure does not violate the Administrative Procedure
      Act, and it does not violate due process of law.

Withrow v. Larkin, 

, 56 (1975).                In his depositions, Dr.

Davis repeatedly testified that he had not prejudged issues, that

his decisions were based on information available at the time, and

that he had decided to refuse MSP’s permit application after an

unbiased review of the evidence involved in the case.                          The fact

that Dr. Davis came to one conclusion based on some evidence did

not   at   all    prevent    him   from   deciding        the   issue   fairly       when

confronted with all the evidence.               See NLRB v. Donnelly Garment

Co., 

, 236-37 (1947) (holding that a hearing examiner’s

prior adverse ruling did not prevent him from adjudicating the same

case on retrial even though the examiner’s initial decision had

been reversed for improper exclusion of evidence).


                                          29
                                       B

     Ms. Sykes was one of the United States' attorneys in the

enforcement action at pre-trial and early in the trial itself.

After Dr. Davis and permit staff determined initially that EPA

should deny MSP's permit application and identified the grounds for

that determination, Ms. Sykes wrote a draft of the findings of fact

and conclusions of law that served as the basis for EPA's statement

justifying the permit denial.      Permit staff official Elaine Taylor

provided evidence that Ms. Sykes became involved only after the

permitting branch had completed an exhaustive review of MSP’s six

volume   application,    after   Dr.       Davis   had   approved   the   staff

recommendation to deny the permit, after the staff had identified

the grounds for the refusal, and several months after Ms. Sykes

resigned from the district court litigation team.              EPA relied on

Ms. Sykes because of the illness of another attorney.

     Ms. Sykes’ role in the process was less than ideal, and the

EAB correctly labeled it a mistake in judgment.            In re Marine Shale

Processors, Inc., Dkt. No. 06900009, RCRA Appeal No. 94-12, 

, at * 23 (EPA 1994).      The question, however, is whether Ms.

Sykes’ role denied MSP due process.           We think not.

     Ms. Sykes “is entitled to the normal presumption of good faith

that, in courts of law, government officials still enjoy, that must

be refuted by well-nigh irrefragable proof.”               Starr v. FAA, 

, 315 (7th Cir. 1979); see Schweiker v. McClure, 

, 195-96 (1982).     MSP’s burden is to persuade us that the use of

Ms. Sykes posed “such a risk of actual bias or prejudgment that the


                                    30
practice must be forbidden if the guarantee of due process is to be

adequately implemented.”   

.   In this case,

Ms. Sykes’ made no decision at all.     She had no power to decide

whether to grant MSP its permit, nor did she have power over those

making that decision.   Her role consisted entirely of articulating

the thoughts and decisions of others.    Even if Sykes’ mind were

“irrevocably closed,” FTC v. Cement Institute, 

, 701

(1948), she had a small role in the decision making process.

     MSP analogizes Sykes’ role to that of a federal court law

clerk and argues that our decision in Hall v. SBA, 

(5th Cir. 1983), mandates reversal here.    In Hall, we remanded a

judgment for a new trial because a magistrate judge used a law

clerk who from previous experience possessed intimate knowledge of

the facts of the case and who had previously written a letter to

the defendants stating that she agreed with the plaintiff.     Hall

does not control our decision in this case for two reasons.   First,

Hall was a decision under 28 U.S.C. § 455, which governs judicial

conduct.   “As this and several other circuits have recognized,

section 455 establishes a statutory disqualification standard more

demanding than that required by the Due Process Clause.”      United

States v. Couch, 

, 81 (5th Cir. 1990); see also Dirt,

Inc. v. Mobile County Commission, 

, 1566 (11th Cir.

1984) (“Although such an appearance of bias is clearly present in

this case, the standards governing administrative proceedings are

far more relaxed than those controlling judicial hearings.”).

Second, the Hall law clerk was involved throughout the entirety of


                                 31
the trial; she wrote bench memoranda, administered the case until

the end of litigation, and had daily informal access to the

magistrate. The danger existed that her bias affected the decision

itself.     In this case, the record shows that EPA used Sykes only

after unbiased staff had reached the tentative decision to deny

MSP’s permit and had identified the grounds for the denial, and

that Sykes had no role in the process beyond providing a draft of

the eventual findings of fact and conclusions of law.

       By comparison, we note that district courts occasionally adopt

wholesale the findings of fact and conclusions of law written by a

victorious litigant.      While we discourage this practice, we have

never radically altered the standard of review in such cases, much

less concluded that such an adoption results in a per se due

process violation.    See Lewis v. NLRB, 

, 1272 n.5 (5th

Cir. 1985); James v. Stockham Valves & Fittings Co., 

,

314 n.1 (5th Cir. 1977) cert. denied, 

(1978).                  We

tolerate the occasional use of this device because of our trust

that district courts will closely examine the proposed findings and

will   carefully   consider    the   objections      and   arguments   of   the

opposing party.    In this case, EPA formalized exactly this sort of

review process.    After Sykes wrote the proposed findings, EPA gave

MSP an opportunity to criticize her work.            MSP took full advantage

of   this   opportunity   by   filing   18   boxes    of   argument.    EPA’s

permitting staff then reexamined the findings and adhered.

       These facts distinguish this case from Bethlehem Steel Corp.

v. EPA, 

, 1009 (7th Cir. 1980), upon which MSP


                                     32
principally relies.      MSP does not allege that EPA failed to

disclose the grounds for its decision, that prosecutorial staff

sought to delay the timing of an adjudication in order to gain a

tactical advantage, or that adjudicatory staff sought to force MSP

to waive certain litigation defenses in return for favorable

treatment on its permit application.           It was the combination of all

of these factors, together with the improper mixing of adjudicatory

and prosecutorial staff, that concerned the Seventh Circuit in

Bethlehem Steel.

     We question whether the use of Sykes as a federal law clerk

would have passed the muster under section 455, given our statement

in Hall that “[e]very judge has suffered a change of heart after

reaching a tentative decision.               Much might happen during the

research and writing to affect the decision.             Until the decision

was signed and rendered, it was in pectore judicis, subject to

possible   

.   Nevertheless,      the

constitutional     standard    for    agency      adjudication   is   not    as

stringent, and we hold that EPA provided MSP due process of law in

its review of the permit application.



                                      IV

     MSP makes one final argument.            It contends that EPA exceeded

its statutory authority by basing the permit denial in part on

MSP's poor history of compliance with environmental laws.                   We

refuse to reach this argument.              We have upheld EPA’s denial of

MSP’s permit on other grounds, and MSP does not argue that the


                                      33
inclusion of this alternative grounds for decision renders the

permit denial infirm.

     AFFIRMED.




                              34