The individual plaintiffs in this case are present or former civilian employees of the Department of the Army employed at the Rock Island Arsenal ("Rock Island") in Rock Island, Illinois. They, along with their labor union, American Federation of Government Employees, Local 2119, brought this action against the federal defendants,1 who are officials and officers of the United States Government. The plaintiffs contend that the Army violated the provisions of the Arsenal Act, 10 U.S.C. sec. 4532, with regard to two defense programs: (1) the production of tank gun mounts for the upgraded M1A2 Abrams tank and (2) the development of the Lightweight 155mm ("LW155") Towed Howitzer. The plaintiffs allege that the Army violated the Act by allowing these items to be produced by private contractors without first undertaking a cost analysis, required by the Act ("Arsenal Act cost analysis"), to demonstrate that production of those items could not instead be performed at a government-owned facility on an economical basis. The district court granted summary judgment to the defendants. For the reasons set forth in the following opinion, we affirm the judgment of the district court.

* BACKGROUND

Rock Island is a government-owned weapons manufacturing facility, one of only two remaining manufacturing arsenals operated by the Army today. It has produced a number of different forms of weaponry for many years, including both of the items at issue in this case--gun mounts and howitzers. However, in recent years the number of manufacturing projects performed at Rock Island has de clined. As a result, the number of civilian workers employed at Rock Island has been reduced continually, and some employees have been placed involuntarily in lower-graded positions.

Since 1976, the Army has acquired and upgraded the M1 Abrams series of tanks from Chrysler Defense, Inc. and its corporate successor General Dynamics Land Systems ("GDLS"). One component of this tank and its upgraded versions is a gun mount. Prior to 1982, all of the gun mounts for the Abrams tank system were produced at the Detroit Army Tank Plant ("DATP") by the contractor.2 In 1982, due to increased demand for the tank, the Army and GDLS amended their contract to provide that one half of the gun mounts needed for the tank would be provided separately to GDLS as government- furnished material, while the other half would continue to be made by GDLS personnel at the DATP ("the DATP gun mounts"). Since that time, the gun mounts produced by the government have been manufactured at Rock Island. The contract for the M1A2 tank upgrade program, at issue in this case, provides that GDLS is responsible for delivering the tank system in its entirety and for determining where it will produce the various components of the system. It also states that the gun mounts needed for the M1A2 upgrade will be produced in the manner described above--half by the government at Rock Island and half by GDLS at the DATP.

In 1995, pursuant to the Defense Base Closure and Realignment Act of 1990, 10 U.S.C. sec. 2687, a base closure commission recommended to Congress that the DATP be closed. At this time, the Army discussed the possibility of having all gun mount production for the M1A2 upgrade program transferred to a GDLS- owned-and-operated facility. Ultimately, in 1996, the Army and GDLS amended their contract to allow for the transfer of equipment used toproduce the DATP gun mounts to a GDLS owned-and-operated facility in Muskegon, Michigan. Production of those gun mounts began at the Muskegon site in March 1997. In effecting this site change, the Army determined that a decision to transfer production on this half of the gun mounts was not subject to an Arsenal Act cost analysis. It took the view that the Act applied only to the manufacture of the "end-item" of a contract and not to the production of each component of that end- item. Because the end-item of the contract for the upgraded M1A2 tank system (the tanks themselves) was being made at a government factory, the Lima Army Tank Plant, the Army determined that a cost analysis was not required.3

Production of the other half of the gun mounts was not shifted from Rock Island to a GDLS facility. Because those items were acquired by the Army directly from a government-run arsenal as a separate item of supply, the government determined that any decision to remove them from Rock Island would require an Arsenal Act cost analysis. After a cost analysis showed that it would be eighteen percent cheaper to continue to produce these gun mounts at Rock Island, as opposed to at a GDLS facility, the Army did not shift their production.

In December 1994, the Marine Corps and the Army collaborated to begin work on the LW155 Towed Howitzer program. This program was to produce a new design for the LW155 Towed Howitzer to replace the previous model, which had tactical and physical shortcomings. Private contractors had been developing prototypes of a lighter-weight howitzer for a number of years at their own expense. With the LW155 howitzer program, the Marine Corps and the Army hoped to take advantage of the technology being developed by these private contractors in order to save money and time on research and development. As a result, the Army's Armament Research, Development and Engineering Center announced a market survey to identify a source for the new LW155 howitzers and later issued draft solicitations to obtain comments from industry on future production of the howitzers.

A formal request for proposals to obtain the LW155 howitzer contract was issued in April 1996. It was contemplated that this contract would be awarded after a two- step process. First, prospective offerors would deliver a prototype howitzer for initial screening and technical documentation, with each eligible offeror to receive a fixed-price contract to support further development and testing. Second, the government would hold a "shoot-off" between all of the offerors' howitzers; the winner of this shoot-off would receive a contract for the engineering and development phase of the program and, after that phase was successfully completed, a full-scale production contract.4 This process went forward, and following the shoot-off, the government entered into an engineering and manufacturing contract with Cadillac Gage Textron, Inc. ("Textron").5

It was anticipated that a total of 869 howitzers would be produced under the contract. Funding for the howitzer program initially was provided by the Marine Corps, although both services will contribute financially to its budget. The Army is also developing a "Towed Artillery Digitalization" ("TAD") system, a location and targeting system that will be included as part of the LW155 howitzer. In the course of its participation in the LW155 howitzer program, the Army did not engage in an Arsenal Act cost analysis to determine whether the howitzers instead could be produced more economically at Rock Island.

On March 5, 1997, the plaintiffs brought this action in the district court for declaratory and injunctive relief under the Administrative Procedure Act ("APA"). They claimed that the defendants had violated a number of statutes, including the Arsenal Act, with regard to the production of both the DATP gun mounts and the LW155 howitzers. The district court dismissed the action for lack of standing. We affirmed in part and reversed and remanded in part, holding that the plaintiffs had standing to sue only under the Arsenal Act. See Am. Fed'n of Gov't Employees, Local 2119 v. Cohen, 171 F.3d 460 (7th Cir. 1999).

Following remand and the filing of a 4,025 page administrative record, the case was submitted to the district court on motions for summary judgment filed by the plaintiffs and the defendants. On August 23, 2000, the court granted the defendants' motion for summary judgment. The court first examined the Arsenal Act. It concluded that the statute's language, which states that the Army "shall have supplies . . . made" in Army facilities if it is economical to do so, 10 U.S.C. sec. 4532(a), was mandatory, rather thanpermissive in its direction. The court also concluded that what constituted "supplies" covered by the Act was less clear. Consequently, the court concluded, the Army's construction of that ambiguous term was entitled to deference under Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984).

The district court then examined whether the Arsenal Act's cost-analysis requirement applied to the production of the DATP gun mounts. The court determined that the Act's provisions did not apply for two reasons. First, it noted the determination of the Army and the Comptroller General that the Arsenal Act does not apply to contracts that are justified on an independent statutory basis. It then explained that, because the DATP gun mounts were produced as part of a "sole source" contract, a type of contract governed by a provision in the Competition and Contracting Act of 1984 ("CICA"), the Arsenal Act's cost-analysis requirement did not apply to those items. Second, the court agreed with the Army's rationale that the Arsenal Act only governed determinations regarding the production of the ultimate end-item contracted for, not the production of every component part of a larger system. Therefore, production of the DATP gun mounts, a component of the larger contract involving the upgrade of the M1A2 Abrams tank, was not subject to an Arsenal Act cost analysis.6

The court also agreed with the defendants that the Arsenal Act did not apply to the LW155 howitzer program. Because the evidence demonstrated that the LW155 howitzer program was a Marine Corps program to be administered under the Department of the Navy's acquisition regulations, the court determined that the Army's Arsenal Act did not apply to contracting decisions for that program.

The district court therefore awarded summary judgment to the defendants. The plaintiffs now appeal that decision to this court.

The plaintiffs submit that the district court erred in determining that the defendants were not required to engage in an Arsenal Act cost analysis before allowing the DATP gun mounts and the LW155 howitzers to be produced by private contractors. After delineating the proper standard of review, we shall address both of the plaintiffs' contentions.

Our review of the district court's decision to grant summary judgment is de novo. See Mosher v. Dollar Tree Stores, Inc., 240 F.3d 662, 666 (7th Cir. 2001). Summary judgment is appropriate when "the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law." Fed. R. Civ. P. 56(c); see also Wade v. Lerner New York, Inc., 243 F.3d 319, 321 (7th Cir. 2001). Therefore, in reviewing a grant of summary judgment, like the district court, we must view all of the facts and draw all reasonable inferences in favor of the nonmoving party. See Del Raso v. United States, 244 F.3d 567, 570 (7th Cir. 2001); Basith v. Cook County, 241 F.3d 919, 926 (7th Cir. 2001).

The district court's interpretation of the Arsenal Act's terms is a conclusion of law that we review de novo. See Ulichny v. Merton Cmty. Sch. Dist., 249 F.3d 686, 699 (7th Cir. 2001). The district court determined that, because it was reviewing the Army's interpretation of a statute that the Army was charged with administering, it would defer to that Department's position on these issues under the holding of Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 842-43 (1984). In Chevron, the Supreme Court explained that, if the intent of a statute is clear, the court and the agency must give effect to the unambiguously expressed will of Congress. See id. at 842-43. However, when the statute is silent or ambiguous with respect to an issue before the court, the court should defer to the agency's interpretation so long as it is based on a permissible construction of the statute. See id. at 843-44.

As the Supreme Court has clarified recently, Chevron deference only applies when "it appears that Congress delegated authority to the agency generally to make rules carrying the force of law, and that the agency interpretation claiming deference was promulgated in the exercise of that authority." United States v. Mead Corp., 121 S. Ct. 2164, 2171 (2001). Such delegation will often be demonstrated by an agency's power to engage in adjudication or notice-and-comment rulemaking that produces regulations or rulings for which deference is claimed. See id.; Christensen v. Harris County, 529 U.S. 576, 587 (2000). In contrast, informal agency interpretations such as those contained in "opinion letters . . . policy statements, agency manuals, and enforcement guidelines, all of which lack the force of law--do not warrant Chevron- style deference." Christensen, 529 U.S. at 587; see also Mead, 121 S. Ct. at 2175. Nevertheless, given the specialized experience and broader information available to such an agency, these informal interpretations are "entitled to respect" to the extent that they have the "power to persuade." Christensen, 529 U.S. at 587; see also Mead, 121 S. Ct. at 2175-76; Skidmore v. Swift & Co., 323 U.S. 134, 139-40 (1944).

This case requires us to examine the text and meaning of the Arsenal Act,7 which today reads as follows:

(a) The Secretary of the Army shall have supplies needed for the Department of the Army made in factories or arsenals owned by the United States, so far as those factories or arsenals can make those supplies on an economical basis.

(b) The Secretary may abolish any United States arsenal that he considers unnecessary.

The basic purpose of this statute is well-recognized. As we noted when this case was first before this court, the statute is "aimed at preserving the government's in-house military production capabilities." Cohen, 171 F.3d at 473. We further noted that the statute's legislative history explained that "'[t]he purpose of [the Act] is to compel the executive officers of the Government to have Government work done at such arsenals . . . and to cease handing out appropriations to private manufacturers.'" Id. at 474 (quoting 59 Cong. Rec. 4157 (March 10, 1920) (remarks of Congressman Sanford)). A December 15, 1960, opinion letter of the Comptroller General (File No. B-143232) also explains that "the basic concept of the statute would appear to be a requirement that Government-owned industrial facilities should not be permitted to lie idle if it would be possible to use such facilities [to produce supplies] at a cost to the Government no greater than the cost of procuring such needs from private industry." R.50, Ex.18 at 5. Therefore, the statute requires that, before the Army makes a decision to produce supplies at a non-governmental facility, it must engage in "a comparison of all costs incurred by the Government as a result of producing an article in Government-owned facilities, with the price at which the article could be purchased from a private manufacturer." Id. If that cost comparison demonstrates that it is more economical to produce the supplies at a government factory or arsenal, then the Act dictates that the Army must do so.

However, as the district court noted, the Act is ambiguous as to what items constitute supplies. Moreover, the Act's text gives no hint as to how its terms should be interpreted when they conflict with other military procurement statutes or regulations.

We now turn to whether the Arsenal Act governs the production of the DATP gun mounts, now manufactured at GDLS' Muskegon facility, that are a part of the M1A2 Abrams tank upgrade program. The district court held that the Arsenal Act's provisions did not apply to the production of these items because they were manufactured as part of a sole source contract permitted as an exception to standard military procurement procedures as set forth in the CICA.

The CICA requires executive agencies to "obtain full and open competition through the use of competitive procedures" when undertaking the procurement of property or services. 10 U.S.C. sec. 2304(a)(1)(A). We have noted that the CICA's purpose is to "save money, curb cost growth, promote innovation and the development of high quality technology and to maintain 'the integrity of the expenditure of public funds.'" Cohen, 171 F.3d at 472 (quoting S. Rep. No. 98-50, at 2-4 (1984), reprinted in 1984 U.S.C.C.A.N. 697, 2174, 2175). However, the CICA lists a number of situations in which the armed forces are not required to follow the procedures for competition set forth in the Act. See 10 U.S.C. sec. 2304(c)(1).9 Among those exceptions is the sole source exception at issue in this case, which states that:The head of an agency may use procedures other than competitive procedures only when . . . (1) the property or services needed by the agency are available from only one responsible source or only from a limited number of responsible sources and no other type of property or services will satisfy the needs of the agency[.]

10 U.S.C. sec. 2304(c). There is no dispute that this provision was the basis for the Army's determination to grant GDLS the contract to complete the M1A2 Abrams tank upgrade program.

The plaintiffs contend that the district court erred in determining that a contract that is made under the CICA's sole source exception is not subject to the Arsenal Act's mandates. However, after reviewing the Army's interpretation of the statute, as confirmed by the legislative history of the Arsenal Act and an authoritative interpretation of the Act's meaning by the Comptroller General, we believe that the district court correctly held that the Act is inapplicable in such a case.

Although the Army's interpretation of this statutory provision does not appear in a source deserving of Chevron deference,10 we nevertheless ought to study the Army's interpretation of the statute that it administers and, to the extent that it has the "power to persuade," it is "entitled to respect." Christensen, 529 U.S. at 587; see also Mead, 121 S. Ct. at 2175-76; Skidmore, 323 U.S. at 139-40.

The record establishes that the Army consistently has interpreted the Arsenal Act's provisions to be inapplicable to contracts reached under the exceptions listed in CICA's Section 2304(c). For example, a July 20, 1992, Army memorandum entitled "Implementation of Statutory Authorities for Manufacturing by Army Industrial Facilities" states that the Arsenal Act's provisions do not apply to "Army supply requirements under solicitations restricted to particular sources." Administrative Record ("A.R.") at 2775. Therefore, the Arsenal Act "does not affect the ability to award supply contracts to particular suppliers . . . to maintain their availability in the event of national emergency or industrial mobilization [the exception listed in the CICA's Section 2304(c)(3)(A)]." Id. Although noting that the Arsenal Act typically requires the Army to manufacture supplies at in-house facilities if economical to do so, the memorandum concludes that the "[u]se of CICA . . . excepts requirements from solicitation under the Arsenal Statute." Id. Similarly, a June 6, 1995, draft Army document entitled "Regulation on Manufacturing and Subcontracting by Army Industrial Facilities" lists certain categories of items that are not appropriate for a "make or buy evaluation" (a term synonymous with the Arsenal Act's cost-analysis requirement). Id. at 2849, 2853. Among these categories are included "items that are restricted to [a] particular source of supply pursuant to a statutory exception to the requirement for full and open competition." Id. at 2853. Moreover, a November 6, 1995, memorandum from Army Material Command Associate Counsel Paul D. Harrington, regarding the application of the Arsenal Act to the production of the DATP gun mounts, states that an Arsenal Act cost analysis would "not apply if an award to GDLS is justified on an independent statutory basis, such as a properly applicable exception to full and open competition under the Competition in Contracting Act, 10 U.S.C. Sec. 2304." Id. at 1379.11

This position of the Army finds support in other material that sheds light on the relationship of the Arsenal Act to the other statutes regulating military procurement. First, the legislative history of the Arsenal Act suggests that it was not meant to apply to contracts such as those based on the sole source exception. When the statutory language that largely mirrors that of the present Arsenal Act was added as Section 101(e) of the Army Organization Act of 1950, Congress clarified its intent that:

Subsection[ ] 101 . . . (e) . . . do[es] not replace or modify the provisions of the Armed Services Procurement Act of 1947 ["ASPA"] . . . which act applies uniformly to all of the armed services [and] vests authority in matters pertaining to contracts for the procurement of materials and services in the heads of the three military departments . . . .

H.R. Rep. No. 81-2110, at 13 (1950), reprinted in 1950 U.S.C.C.A.N. 2607, 2620. As we have noted, although the ASPA did not contain the sole source exception at that time, it did include a number of exceptions to competitive procurement procedures that are like those presently found in Section 2304(c) of the CICA. This legislative history confirms that Congress believed that the Arsenal Act's requirements might at times conflict with those imposed on all the armed forces by the ASPA, and that, when such a conflict arose, the strictures of the Arsenal Act should not hinder the ASPA's application.

More significantly, the defendants' position also is bolstered by the December 15, 1960, opinion letter of the Comptroller General. The district court described this opinion letter as the "lead decision interpreting the requirements of the Arsenal Act," R.74 at 13 n.3 (quotation marks and citation omitted), and even the plaintiffs refer to it as the "landmark analysis of the Arsenal Act," Appellants' Br. at 31. The opinion letter explains that the Arsenal Act's preference for government production will at times conflict with those exceptions in the ASPA that require the government to contract with a private entity without undergoing a competitive bidding process. The opinion letter then concludes that, to the extent that the two statutes do conflict, "the provisions of the Armed Services Procurement Act would control." R.50, Ex.18 at 8.12 As an example of such a situation, the opinion letter cites the ASPA's Section 2304(a)(16) (now listed in the CICA's Section 2304(c)(3)(A)), "which authorize[s] the Secretary of the Army to negotiate a contract with a particular supplier in the interest of national defense and industrial mobilization, notwithstanding the existence of other private or Government-owned production facilities." Id. The opinion also notes that there might be circumstances in which other exceptions in the ASPA would allow the government to invoke the ability to contract with a private entity "notwithstanding the existence of Government-owned facilities capable of economically producing the product needed." Id. Ultimately, it concludes that "unless a particular procurement of Army supplies falls within an exception prescribed by other law," the Arsenal Act's provisions apply to production decisions regarding those supplies. Id. at 9 (emphasis added).

The plaintiffs argue to the contrary that application of CICA's sole source exception can coexist with the Arsenal Act's mandates, instead of preempting the use of the Arsenal Act's cost-analysis provision altogether. They maintain that the Arsenal Act first requires the Army to make a decision as to whether particular supplies can be produced economically by government personnel. The plaintiffs assert that, only if this cost analysis determines that such work may be undertaken by a private contractor, do the CICA's provisions (regarding how such a contractor should be selected and when competitive procedures should be bypassed) come into play.

The plaintiffs cite no authority for this interpretation, which certainly conflicts with both the legislative history of the Arsenal Act and the Comptroller General's opinion with regard to the exceptions to competitive bidding practices enumerated in CICA's Section 2304(c). These exceptions, including the sole source exception, reference situations in which either a national emergency or a narrow set of compelling circumstances requires that a branch of the armed services must contract with a particular private entity.13 As we have noted, the Comptroller General's opinion cites one such exception, now located in the CICA's Section 2304(c)(3)(A), which allows the Army to negotiate a contract with a private supplier, without regard for competitive practices, when it is necessary to do so to prepare for a national emergency or to achieve industrial mobilization. As in that example, it is not difficult to imagine circumstances in which the Army would need to contract with the sole source of a particular item without regard for the cost analysis required by the Arsenal Act. The Arsenal Act's legislative history, bolstered by the Comptroller General's interpretation, counsels that, in such circumstances, the Army should be able to contract with the private entity without first satisfying the requirements of the Arsenal Act.

In sum, the defendants have pointed to evidence drawn from consistent Army practice, legislative history and a landmark decision of the Comptroller General, all of which suggest that an Army contract entered into pursuant to CICA's sole source exception should not be subject to the Arsenal Act's requirements. Moreover, a reading of the exceptions listed in the CICA's Section 2304(c) make it understandable why this would be so, because these exceptions deal with important circumstances in which it would be either impractical or potentially dangerous to hinder the agency from contracting with a particular private entity. In contrast, the plaintiffs point to no persuasive authority to demonstrate that their alternative reading of the interplay of the two acts is a viable one. Consequently, we believe that the district court correctly granted summary judgment to the defendants with regard to the DATP gun mounts.14

The district court also concluded that the Army did not violate the Arsenal Act with regard to the production of the LW155 howitzers. The court determined that the Act, which applies only to the Army, was not implicated by the howitzer program because that program was directed by the Marine Corps (a component of the United States Naval Service) and was intended to be executed pursuant to Department of the Navy acquisition regulations. We agree with the district court's decision.

The evidence of record persuades us that the LW155 howitzer program is not an Army procurement program and that it was meant to be directed under Navy acquisition regulations. The program is a joint endeavor between the Marine Corps and the Army, to which both services devote personnel and funds and from which both ultimately will receive howitzers for their use. However, the Marine Corps has the clear authority to direct the program's implementation. A November 3, 1995, "Memorandum of Agreement" between the two services regarding the program's direction and scope clearly states:

General Policy. As the Lead Service acting under the guidance of the ASN (RDA) [Assistant Secretary of the Navy (Research, Development and Acquisition)], the U.S. Marine Corps, represented by the COMMARCORSYSCOM [Commander, Marine Corps Services Command], has the authority to direct the program under the policies and procedures set forth in Department of Defense (DOD) and Department of the Navy acquisition regulations. The PEO-FAS [the Army's executive agent for the LW155 howitzer program] will execute the program per the decisions and direction of the COMMARCORSYSCOM and the ASN (RDA).

A.R. at 2789 (emphasis added); see also id. at 2795 ("FY96 Acquisition Plan No. 1-96 for Lightweight 155mm Howitzer (LW155)") (hereinafter "FY96 Acquisition Plan") (stating that "the U.S. Marine Corps . . . has the authority to direct the program under the policies and procedures set forth in Department of Defense and Department of the Navy acquisition regulations"). This memorandum explicitly designates the Marine Corps as the service in charge of the howitzer program and explains that Navy acquisition regulations, not those of the Army, apply to that program. Cf. id. at 2797 (FY96 Acquisition Plan) (stating that "[t]he Marine Corps and the Army signed a Joint Memorandum of Agreement on 3 Nov 1995 which stipulated how the Army would support a Marine Corps lead LW155 program").

Moreover, a number of other factors support the conclusion that the Marine Corps is the service charged with the direction of the howitzer program. One such factor is the division of responsibilities allocated to each service regarding the program's implementation. The Marine Corps' responsibilities as "Lead Service" include retaining authority over all program funds and their transfer as well as competing for necessary resources to support execution of the program. Id. at 2790. In contrast, the Army's responsibilities as the "Participating Service" include the supportive tasks of offering procurement and policy "guidance," providing staff and facilities and executing the Marine Corps' contracting actions. Id. at 2791; see also id. at 2795 (FY96 Acquisition Plan). Additionally, the Marine Corps plays a significantly larger role in providing funding for the howitzers. The Marines contributed all of the initial funding for the program in its first five years and will have contributed $1.085 billion of the $1.853 billion total program cost (58.6%) by 2009, the last year for which cost evaluations were reflected in the record.15 Lastly, the Marines also will receive a far greater number of the howitzers than the Army. Of the 869 total howitzers that will be produced under the program, the Marines will be allotted 450 while the Army will receive only 273.16

The Arsenal Act states that the "Secretary of the Army shall have supplies needed for the Department of the Army made in factories or arsenals owned by the United States." 10 U.S.C. sec. 4532. This language suggests that, for supplies to fall within the purview of that Act, the Secretary of the Army must have the authority to direct that their production follow Arsenal Act regulations. However, in this case, the documents surrounding the howitzer program demonstrate that the Navy was the designated service that would retain this power, not the Army. The plaintiffs do not explain how, after ceding authority for acquisitions under the howitzer program to the Marine Corps, the Army could then force the Marine Corps to abide by its acquisition policies.

For these reasons, we believe that the district court correctly determined that production of the LW155 howitzers was not subject to an Arsenal Act cost analysis.

The district court did not err in determining that the decisions to produce the DATP gun mounts and the LW155 howitzers at private facilities were not subject to the Arsenal Act's requirements. Accordingly, we affirm the judgment of the district court.