524 F.3d 627
5th Cir.
2008

Lead Opinion

SOUTHWICK, Circuit Judge:

A boat manufacturer, which is the defendant in a personal injury suit, appeals the district court’s dismissal of its third-party complaint against the United States. This defendant sought indemnity on the basis of alleged negligence by an Army recreational facility, which owned the boat and rented it to an off-duty soldier. The manufacturer argues that jurisdiction exists because the service member who was injured on the boat was not engaged in activity incident to military service. We agree, reverse and remand.

FACTS

On Saturday, April 16, 2005, near noon, the Plaintiff, Daniel J. Regan, met his friend, John Vandergriff, at the Army’s Toledo Bend Morale, Welfare, and Recreation (MWR) facility.1 The recreational facility is located forty-five miles from the military base at Fort Polk, Louisiana. Both Regan and Vandergriff were active duty Army staff sergeants assigned to Fort Polk who were off duty for the day. Earlier that morning, Staff Sergeant Van-dergriff had rented a pontoon boat from the MWR facility. Staff Sergeant Regan was a Missouri National Guardsman who had been on active military duty with the Army for more than a year. Regan’s civilian girlfriend was with him. Including the other soldiers and guests who joined them, eight people were on the pontoon boat. The group spent a day relaxing, swimming, and drinking.

The MWR rental boats are kept in a small cove that is part of the Army recreational facility, but the boats may be taken out into the Toledo Bend Reservoir. At approximately 7:00 p.m. that day, the pontoon boat was in the waters of the Reservoir. As Regan stood up from his seat in the bow of the boat, Vandergriff reduced the speed of the boat, causing Regan to stumble forward. Regan grabbed the front gate in effort to regain his balance and prevent him from falling overboard. The gate ripped from its post, causing Regan to fall off the front of the boat. The pontoon boat’s propeller struck Re-gan’s right leg, causing serious injuries and ultimately leading to the leg’s amputation.

After the accident, the Army conducted an investigation. The resulting report determined that the soldiers’ activities were not mission or training related, that the accident was off-post, and that Vandergriff and Regan were off duty at the time of the accident. The report found that neither Vandergriff nor Regan had received the necessary license to operate the pontoon boat on the Reservoir, and the pontoon boat was civilian, not military, equipment. The report included recommendations for improved safety briefings for soldiers renting boats. A different Army report made other recommendations, such as improving safety briefings, equipping each boat with propeller guards, and strengthening the gates at the front end of the pontoons.

The Toledo Bend MWR facility falls under the direct command of the Garrison Commander at Fort Polk, but the facility is run by civilian employees. The facility is open only to military personnel and their dependents, federal civilian employees of Fort Polk, military retirees, and guests accompanying authorized patrons. See Army Reg. 215-1, ¶ 7-1.

Starcraft Marine LLC designed and manufactured the boat Vandergriff rented from the MWR facility. Regan sued Star-craft, two insurance companies, and Van-dergriff in Louisiana state court. The allegations against Starcraft were defective design, manufacture, and marketing of the pontoon boat. Starcraft removed the case to federal court, alleging federal question jurisdiction because the accident allegedly occurred on property over which the United States had exclusive jurisdiction. The district court remanded because Starcraft failed to offer evidence of federal authority over the site of the accident.

Following remand, Starcraft filed a third-party complaint against the United States, alleging claims under the Federal Tort Claims Act (FTCA). The third-party complaint alleged that the United States was negligent in (1) renting the boat to Vandergriff when it was in disrepair; (2) renting the boat to Vandergriff to use in a manner inconsistent with the intended function of the boat; (3) failing to ascertain how Vandergriff and his boating party intended to use the boat; (4) failing to provide adequate instructions to Vandergriff and his boating party regarding the proper use of the boat; and (5) failing to maintain and repair the boat properly.

The United States again removed the case to federal court. After removal, it sought to dismiss the complaint due to what is referred to as the Feres doctrine, which bars tort suits against the United States by or on behalf of service members2 whose injuries arise out of activity incident to their military service. Feres v. United States, 340 U.S. 135, 146, 71 S.Ct. 153, 95 L.Ed. 152 (1950). On March 26, 2007, the district court issued two orders. The first order dismissed the third-party claim against the United States for lack of subject matter jurisdiction under the Feres doctrine. The second order remanded the state-law claims to state court. Starcraft timely appealed from the district court’s judgment. Starcraft presents for our review only the district court’s dismissal of the third-party claim against the United States pursuant to the Feres doctrine.

DISCUSSION

I. Jurisdiction

The order which is the subject of this appeal is the dismissal of the United States as a party. The district court in the same judgment remanded the remaining issues of the litigation to the state court from which the suit had earlier been removed. Whether the dismissal order is a proper subject of appeal is our first legal issue.

Generally, a federal appellate court may not review a remand by a district court to state court for lack of subject matter jurisdiction. 28 U.S.C. § 1447(d). However, when the trial court “clearly and affirmatively” states that it is remanding on a ground other than a lack of subject matter jurisdiction, the Section 1447(d) bar to appeal does not apply. Tillman v. CSX Transportation, 929 F.2d 1023, 1027 (5th Cir.1991). The district court here dismissed all claims over which original jurisdiction existed, then exercised its discretion not to retain supplemental jurisdiction over remaining claims. See 28 U.S.C. § 1367(c). This Circuit has held that a clear and affirmative use of the Section 1367(c) discretion not to retain supplemental jurisdiction but instead to remand, is reviewable on appeal for an abuse of discretion. Thomas v. LTV Corp., 39 F.3d 611, 616 (5th Cir.1994). We note that the Supreme Court has recently found it “far from clear” that such a remand is free of the Section 1447(d) bar. Powerex Corp. v. Reliant Energy Servs., — U.S. -, 127 S.Ct. 2411, 2418-19, 168 L.Ed.2d 112 (2007). The concurring opinion recommends asserting our authority to review the remand order and logically describes some benefits of doing so. However, Star-craft stated in its brief that “it is not appealing the order of remand.” We decline to proceed beyond the appellant’s disclaimer.

Our conclusion about the order appealed here, explained in some detail below, is that the dismissal of the United States was improper. Should our decision suggest the desirability of again seeking to bring all the claims together in one court, a possibility significantly impacted by whatever has occurred in state court since the remand over one year ago, the parties may address that issue in the state and federal trial courts.

Because the district court coupled its dismissal with a remand order, we must answer two questions to determine whether the dismissal order may be reviewed on appeal. The first is whether there is an order that is “distinct and separable” from the remand order, and therefore not encompassed within Section 1447(d)’s bar to review of a remand. First Nat’l Bank v. Genina Marine Servs., Inc., 136 F.3d 391, 394 (5th Cir.1998). The second question is whether that order is either final and ap-pealable under 28 U.S.C. § 1291, or subject to appeal because of some doctrinal exception to finality. Doleac ex rel. Doleac v. Michalson, 264 F.3d 470, 478, 489 (5th Cir.2001).

Turning to the first question, “we may review any aspect of a judgment containing a remand order that is distinct and separable from the remand proper. An order is ‘separable’ if it precedes the remand order ‘in logic and in fact’ and is ‘conclusive.’ ” Genina Marine, 136 F.3d at 394 (citations and some internal quotation marks omitted). The district court’s dismissal of Starcraft’s third-party claims against the Army preceded the remand order; the dismissal is conclusive because “it will have preclusive effect in the state-court litigation and will not be subject to review there.” Id.; see City of Waco v. United States Fidelity & Guar. Co., 293 U.S. 140, 143, 55 S.Ct. 6, 79 L.Ed. 244 (1934). Judge Barksdale in Doleac described the separableness concept as being “rooted” in the City of Waco decision. Doleac, 264 F.3d at 479. A separate order for these purposes does not just determine the forum in which a claim will be heard. It is one that determines whether there is any claim to be heard in any forum. Id. at 487-88. The result of the district court’s dismissal here has unreviewably determined, unless there is review now, that there was no claim to be heard anywhere against the United States on these facts. This is a separate order.

We turn to the second question, which is whether there is some procedural mechanism for an appeal of this order. The usual appeal process is from a final decision of a district court. See 28 U.S.C. § 1291 (circuit courts “have jurisdiction of appeals from all final decisions” of district courts). There can be room for doubt on the application of this statute: “a decision is ordinarily considered final and appeal-able under § 1291 only if it ends the litigation on the merits and leaves nothing for the court to do but execute the judgment.” Quackenbush v. Allstate Ins. Co., 517 U.S. 706, 712, 116 S.Ct. 1712, 135 L.Ed.2d 1 (1996).

The Quackenbush description of finality is consistent with the Supreme Court’s long-standing position that Section 1291 must be given a “practical rather than a technical construction.” Cohen v. Beneficial Indus. Loan Corp., 337 U.S. 541, 546, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949). In Cohen, the Court held that courts could entertain appeals of certain decisions that — although not final in a technical sense — “fall into that small class which finally determine claims of right separable from, and collateral to, rights asserted in the action, too important to be denied review and too independent of the cause itself to require that appellate consideration be deferred until the whole case is adjudicated.” Id.; see Will v. Hallock, 546 U.S. 345, 349-50, 126 S.Ct. 952, 163 L.Ed.2d 836 (2006). This, unsurprisingly, is known as the collateral order doctrine. The doctrine is not an exception to the requirements of a final judgment but instead is “a practical construction of the rule.” Digital Equip. Corp. v. Desktop Direct, Inc., 511 U.S. 863, 867, 114 S.Ct. 1992, 128 L.Ed.2d 842 (1994).

Looking now to our case, we find that the district court entered one final judgment that contained both an order dismissing Starcraft’s claims against the Army and an order remanding all other claims to state court. A similar set of facts appears in a 1998 precedent of this Circuit. Genina Marine, 136 F.3d 391. The initial defendant filed a third-party complaint against a federal entity, the Farmers Home Administration (FmHA), which then removed the suit to federal district court. Once in a federal forum, the FmHA sought to be dismissed based on sovereign immunity. The district court dismissed the FmHA and remanded the remainder of the suit to state court. The dismissal was appealed. Id. at 394. Instead of using Cohen’s collateral order analysis, this Court found that “under City of Waco and its progeny, we have jurisdiction to review the district court’s dismissal” of third-party claims against a federal entity, when the order of dismissal was contained in the same judgment but analytically preceded the remand order. Id. By referring to City of Waco and not to the later Cohen decision that first mentioned the collateral order doctrine, this Court may have been making a clear choice of not using that doctrine.

Here, the dismissal resolved the merits of Starcraft’s third-party claims against the Army. The district court then remanded the remaining claims, leaving no ongoing federal district court litigation upon which this court would “intrude” by entertaining Starcraft’s appeal. See Cohen, 337 U.S. at 546, 69 S.Ct. 1221. We find no compelling need to label our juris dictional conclusion as being under the collateral order corollary to an appeal from the usual final judgment, or whether the corollary is inapplicable and the base rule is being applied. For example, it has been held by the Third Circuit that a similar appeal was justified under both approaches. Carr v. Am. Red Cross, 17 F.3d 671, 675 (3d Cir.1994) (dismissal of one party and remand of remaining litigation was both a final order under Section 1291 and appealable as a collateral order).

We therefore have jurisdiction to review the dismissal. Our sole issue is whether there was subject matter jurisdiction for the third-party claim against the United States. We do not have before us the merits of that claim.

II. The Feres doctrine

Starcraft’s third-party claims against the United States were dismissed because of this doctrine: “the Government is not liable under the FTCA for injuries to servicemen where the injuries arise out of or are in the course of activity incident to service.” Feres, 340 U.S. at 146, 71 S.Ct. 153. This doctrine has been much-criticized. E.g., United States v. Johnson, 481 U.S. 681, 700, 107 S.Ct. 2063, 95 L.Ed.2d 648 (1987) (Scalia, J., dissenting) (lists appellate decisions and academic articles supporting the view that “Feres was wrongly decided and heartily deserves the ‘widespread, almost universal criticism’ it has received”); Deirdre G. Brou, Alternatives to the Judicially Promulgated Feres Doctrine, 192 Mil. L.Rev. 1, 4 (2007) (doctrine overly broad). Regardless, our task is to determine the doctrine’s current reach.

The district court made its determination about the application of Feres to Staff Sergeant Regan’s actions at Toledo Bend. After ruling that there was no jurisdiction over Starcraft’s claims against the United States, it dismissed. Our review of this legal issue is de novo. Meister v. Texas Adjutant General’s Dep’t, 233 F.3d 332, 336 (5th Cir.2000).

We are examining a claim by a third party, a boat manufacturer, who seeks indemnity from the United States for claims that potentially would be barred under Feres if brought against the Government by the service member. The Supreme Court addressed the operation of the Feres doctrine on a defendant’s effort to implead the Government for indemnity or contribution in situations where “the original defendant claims that the United States was wholly or partially responsible for the plaintiffs injury .... ” Stencel Aero Engineering Corp. v. United States, 431 U.S. 666, 669-70, 97 S.Ct. 2054, 52 L.Ed.2d 665 (1977) (quoting United States v. Yellow Cab Co., 340 U.S. 543, 71 S.Ct. 399, 95 L.Ed. 523 (1951)).

Stencel involved a service member injured due to a malfunction in an aircraft life-support system. 431 U.S. at 667-69, 97 S.Ct. 2054. The service member sued both the United States and the Government contractor that manufactured the malfunctioning life-support system according to the Government’s specifications. Stencel, the Government contractor and manufacturer, cross-claimed against the United States. The motions by the United States for dismissal of both the service member’s suit and Stencel’s cross-claim were granted. Id. The Supreme Court found that in suits based on “an injury sustained by a soldier while on duty, the effect of the action upon military discipline is identical whether the suit is brought by the soldier directly or by a third-party.” Id. at 673, 97 S.Ct. 2054. Both the service member’s and the contractor’s trials would “involve second-guessing military orders, and would often require members of the Armed Services to testify in court as to each other’s decisions and actions.” Id. The dismissal of Stencel’s indemnity suit against the United States was therefore affirmed.

Having found the Feres doctrine to be implicated in this action, we start our effort to understand the doctrine by considering three broad rationales that in the past were identified by the Supreme Court as justifying the doctrine. Whether all or any of the rationales must be served before recovery is barred in a particular case is the next question, but these rationales at least explain the Supreme Court’s justifications for the doctrine.

The first rationale has been described in this way:

The relationship between the Government and members of its armed forces is distinctively federal in character. This federal relationship is implicated to the greatest degree when a service member is performing activities incident to his federal service ....

Johnson, 481 U.S. at 688-89, 107 S.Ct. 2063 (footnotes, internal quotation marks, and citations omitted). Similarly, the Court held that the “scope, nature, legal incidents and consequences of the relation between persons in [military] service and the Government are fundamentally derived from federal sources and governed by federal authority.” Feres, 340 U.S. at 143-44, 71 S.Ct. 153 (quoting United States v. Standard Oil Co., 332 U.S. 301, 305-06, 67 S.Ct. 1604, 91 L.Ed. 2067 (1947)). The Court contrasted the centrality of federal law to military issues with the FTCA’s requirement that the law of the state in which injury occurred governs liability under the Act. Feres, 340 U.S. at 142, 71 S.Ct. 153 (citing 28 U.S.C. § 1346(b)).

In addition, the Court supported the first rationale by distinguishing the provision in the FTCA that the United States is liable under that Act “in the same manner and to the same extent as a private individual under like circumstances .... ” 28 U.S.C. § 2674. The unique relationship between a member of one of the armed services and the government means there are no “like circumstances” with which to make a comparison.

The second rationale focused on alternatives to FTCA liability:

Second, the existence of these generous statutory disability and death benefits is an independent reason why the Feres doctrine bars suit for service-related injuries. In Feres, the Court observed that the primary purpose of the FTCA “was to extend a remedy to those who had been without; if it incidentally [benefited] those already well provided for, it appears to have been unintentional.”

Johnson, 481 U.S. at 689-90, 107 S.Ct. 2063 (footnotes and citations omitted). This rationale focuses on other forms of compensation available to members of the military services whose claims were incident to their service, but did not concern itself with differences in the amount of the compensation available under the FTCA and under the relevant military benefits programs.

The third and final rationale identified by the Supreme Court is this:

Third, Feres and its progeny indicate that suits brought by service members against the Government for injuries incurred incident to service are barred by the Feres doctrine because they are the “type[s] of claims that, if generally permitted, would involve the judiciary in sensitive military affairs at the expense of military discipline and effectiveness.”

Id. at 690-91, 107 S.Ct. 2063 (quoting United States v. Shearer, 473 U.S. 52, 59, 105 S.Ct. 3039, 87 L.Ed.2d 38 (1985) (emphasis omitted)). This last rationale is the one most likely to vary in its arguable existence between cases, because the special relationship of service members with the military and the separate benefits programs exist whenever an injury occurs incident to service.

By 1985, the first two rationales were found to be “no longer controlling.” Shearer, 473 U.S. at 58 n. 4, 105 S.Ct. 3039.3 Though the third rationale continues to be invoked, a separate question is whether, before the Feres bar will apply, there must be evidence that allowing a particular suit would in fact intrude into sensitive military affairs. The Supreme Court answered that question in the negative in a decision handed down two months after Johnson:

A test for liability that depends on the extent to which particular suits would call into question military discipline and decisionmaking would itself require judicial inquiry into, and hence intrusion upon, military matters. Whether a case implicates those concerns would often be problematic, raising the prospect of compelled depositions and trial testimony by military officers concerning the details of their military commands. Even putting aside the risk of erroneous judicial conclusions (which would becloud military decisionmaking), the mere process of arriving at correct conclusions would disrupt the military regime. The “incident to service” test, by contrast, provides a line that is relatively clear and that can be discerned with less extensive inquiry into military matters.

United States v. Stanley, 483 U.S. 669, 682-83, 107 S.Ct. 3054, 97 L.Ed.2d 550 (1987).

The Court concluded that discovery procedures such as depositions that would be needed to determine how disruptive a review of the military judgment would be, would themselves be disruptive. Instead, Stanley sought a “bright line” rule of imposing the Feres bar when the injury was incident to service. Discerning that line would require “less extensive inquiry into military matters.” Id. at 683, 107 S.Ct. 3054. It is true that the third Feres rationale includes the “incident to service” concept, but that concept is also the central element in the definition of the doctrine: “the Government is not liable under the Federal Tort Claims Act for injuries to servicemen where the injuries arise out of or are in the course of activity incident to service.” Feres, 340 U.S. at 146, 71 S.Ct. 153. In summary, Stanley is a straightforward application of the incident to service test. The Court found it improper to inquire into applicability of the third rationale in a specific case.

Though the majority in Stanley in 1987 sought to impose an objective rule that did not examine whether specific litigation would interfere with military judgments, discipline, and management, just two years earlier the Supreme Court stated that the doctrine could not be applied using objective tests. In 1985, the Court found the third rationale critical in making a case-by-case analysis: the location of the incident was not nearly as important in deciding whether Feres should apply as was “whether the suit requires the civilian court to second-guess military decisions, and whether the suit might impair essential military discipline.” Shearer, 473 U.S. at 57, 105 S.Ct. 3039 (citations omitted). The claim was that the Army negligently failed to control a soldier who had once been convicted of manslaughter and was known to be dangerous, a failure that led to the killing of another soldier. Id. at 57-58, 105 S.Ct. 3039. The Court concluded that the allegation went directly to the “management” of the military and litigation over that issue would impermissibly involve the courts in reviewing military judgments. Id. at 58-59, 105 S.Ct. 3039.

The clarity with which Shearer stated that the likelihood of disruption of military judgments in a particular case was critical in determining whether to apply the Feres bar, despite the later Stanley decision’s equally clear statement that it was not, may be what has caused this and other Circuits since the date of both precedents to continue to examine the applicability of the rationales in specific cases. See, e.g., Schoemer v. United States, 59 F.3d 26, 28 (5th Cir.1995); Kelly v. Panama Canal Comm’n, 26 F.3d 597, 599 (5th Cir.1994). Certainly, if the claim is based on an injury that was incident to service and if resolving that claim also would demand second-guessing of military judgments, an application of Feres is clear. We will therefore review the applicability of the last Feres rationale to this indemnity claim.

III. Application of the third Feres rationale

Our analysis of how the military’s judgment would be impacted here starts with Starcraft’s third-party complaint. At least three of its allegations of negligence— renting the boat for use in a manner inconsistent with the intended function of the boat, failing to ascertain the boating party’s intended use of the subject boat, and failing to provide adequate instructions regarding the boat’s proper use — question the Government’s choices about supervision and control of the service members using the Toledo Bend MWR services. Those allegations would require examination of how Toledo Bend implemented the rules that the military itself developed. Another allegation is that this boat was not properly inspected and maintained. As to those issues, no known and specific Toledo Bend policy has been asserted.

How these issues implicate the military to some degree can be seen from the fact that the MWR program falls under the direct authority of the Garrison Commander at Fort Polk. Army Regulation 215-1 and the Garrison Commander’s policy set the basic rules for use. Starcraft’s suit might require the Garrison Commander at Fort Polk to explain to a civilian jury some aspects of his judgment regarding the Toledo Bend MWR facility. The military also is implicated because the facility is not open to the general public. All who use it are subject to the Army and Garrison-established rules. Soldiers who violate those rules are subject to military discipline, and soldiers who damage or lose property are subject to financial liability under Army Regulation 735-5. On the other hand, though Toledo Bend is not open to the public, civilians are allowed to use the facility as guests of the military or federal employees. One guest in this incident was Staff Sergeant Regan’s civilian girlfriend.

Military judgment clearly was exercised after the incident. An official Army investigation of the accident occurred. Specific recommendations were made to improve safety briefings for boat rentals and equipping the boats with radios, first aid kits, higher guard rails, stronger gates, and propeller guards.

Certainly, authority over this MWR program is exercised by the military. The Government argues that allowing Star-craft’s third-party complaint to go forward could involve judicial review of military decisions and regulations concerning the MWR program, in particular, the maintenance and repair of the Army-owned boats, radios, and other equipment. Yet use and maintenance of recreational boats cannot reasonably be compared with military judgments of the more traditional sort. The MWR facility is a civilian-run program. The program does not involve a function, training, or operation unique to the military. Only in the most attenuated sense are the judgments to be reviewed the kind of military decisionmaking that courts should not be second-guessing.

We are not saying that because the activity in question was not a uniquely military one, the issues before us are resolved in favor of suit. We are saying that on the question of the applicability of the third rationale for Feres, the fact that the activity in question has a largely civilian character has relevance.

Also not an issue here is interference with military discipline. Nothing in the record indicates any soldier on the boat was disciplined by the Army, and even if that occurred, we discern no disruption of that process by this suit.

We conclude that this is a situation in which the traditional concerns underlying the third rationale for the Feres bar are absent. The Supreme Court in Stanley may have made this review of the three rationales unnecessary. The Court’s opinion focused on the “incident to service” test and exhibited unconcern for the rationales. We now turn to the unambiguously relevant issue of whether these injuries were incident to Staff Sergeant Regan’s military service.

jTV. Test for an activity being incident to service

The injury to Staff Sergeant Regan occurred when he was off duty, off base, and not engaged in any activity with a direct military purpose. The district court observed that this was “a very close question,” but concluded Regan’s activities to be incident to his military service. We agree with the observation but not the conclusion, as we will explain.

This Circuit uses a three-factor analysis for whether a service member’s injury was incident to military service: (1) duty status, (2) site of injury, and (3) activity being performed. Schoemer, 59 F.3d at 28 (citing Parker v. United States, 611 F.2d 1007, 1013-15 (5th Cir.1980)) (this Circuit refers to these factors as the “Parker test”). We will review each factor.

A. Duty status

Duty status has sometimes been described as the most important of the factors for whether an activity was incident to service. Schoemer, 59 F.3d at 28-29. Duty status is key because it reveals the relation between the service member and the military at the time of injury and therefore how truly incident to service it was. For a National Guardsman who has been mobilized and is serving a period of active federal duty as was Staff Sergeant Regan (identically to a traditional soldier and in contrast to a Reservist or National Guardsman who has not been called to active duty), this factor is not concerned with whether the service member is at that time subject to military discipline. That would be the wrong focus, because active duty service members are always subject to discipline for their actions, not just when they are in uniform or performing military duties. 10 U.S.C. § 802(a)(1). Instead, what is relevant about the status of an active duty service member at the time of injury is where that status is on a continuum between performing the tasks of an assigned mission to being on extended leave from duty. See Parker, 611 F.2d at 1013.

A sampling of definitions relevant to off-duty status used by Staff Sergeant Re-gan’s military service, the Army, is instructive:

Annual leave
Leave granted in execution of a command’s leave program, chargeable leave account. Such leave is also called ordinary leave.
Convalescent leave
A period of authorized absence granted to soldiers under medical care for sickness or wounds and not yet fit for duty. It is part of the treatment prescribed for recuperation and convalescence and is not chargeable as leave.
Emergency leave
Chargeable leave granted for a personal or family emergency requiring the soldier’s presence.
Leave
Authorized absence from place of duty, chargeable against the soldier’s leave account. It is earned at the rate of 2 1/2 days of leave per month for active duty of 30 consecutive days or more, except for periods in nonpay status.
Pass
An authorized absence granted for short periods to provide respite from the working environment or for other specific reasons. At the end of the pass period, soldiers will be at their places of duty or in the locations from which they regularly commute to work.

Army Reg. 600-8-10, Glossary, at 111-12. Though a “furlough” is identified in some caselaw as the military status least likely to evoke a Feres bar, that word is not in this Regulation as current Army terminology.4 A period of leave of more than three days is the present-day equivalent, while the shorter period encompassed within a “pass” would not be quite as distant from the Feres bar.

These varying categories of military leave and passes serve different purposes, but the Army recognizes that its “leave policies are an important command requirement,” and further, that the “use of leave will make a positive contribution to morale, level of performance, and career motivation.” Army Reg. 600-8-10, ¶ 2-1. We note these provisions because they articulate the military purpose served by leave. The purpose certainly is not as closely associated with the core military mission as, say, being on watch in a guard tower at a base in Iraq, but leave serves a military purpose nonetheless. This regulation reinforces the point that duty status is a continuum.

In determining Staff Sergeant Regan’s status when injured, we start with the fact that he was a Missouri National Guardsman who had been mobilized. According to an affidavit in the record, he was on orders calling him to active duty beginning on November 13, 2003, and was scheduled to be released from active duty about a month after this accident.5 He therefore was an active duty Army soldier at the time of the injury.

An active duty soldier, though, will at various times be in different locations on the continuum from being fully engaged in a military mission to being on an extended period of leave. To determine where Staff Sergeant Regan should be placed on that continuum, we start with the official investigation report on the accident. It described Regan simply as being “off duty.”

In his deposition, Regan indicated that he did not need to take leave or apparently even receive a pass to pick up his girlfriend at the Shreveport, Louisiana, airport, and proceed to Toledo Bend where the accident occurred the next day. According to the deposition, the Garrison Commander allowed soldiers to travel a certain mileage distance from Fort Polk, within which Shreveport was included, without taking leave. As we understand his statements, Regan needed neither a pass nor leave because he had no military duties on the days that he was involved in this recreation, and because he remained in sufficient proximity to Fort Polk during his activities.

Though a soldier may be granted a pass for not more than three days (Army Reg. 600-8-10, ¶ 5-27e.(2)), even that form of permission apparently had not been granted to Regan. Unless authorized to be absent, a “soldier remains in an available-for-duty status during normal off-duty hours,” which may be the best explanation of Staff Sergeant Regan’s status when he was injured. Army Reg. 600-8-10, ¶ 5-27c. He apparently was off duty, not technically on leave or pass, and therefore available for duty if called.

We have noted that duty status is considered a critical factor. The facts here are similar to a precedent in which an Army captain, off duty for the weekend, was killed when the mast of the catamaran he was sailing through the Panama .Canal struck hanging electrical wires. Kelly, 26 F.3d at 599. The officer’s widow sued the Panama Canal Commission, which was the agency responsible for the wires. This Court found that Captain Kelly’s “duty status falls along the middle of the spectrum and is not a strong indicator of whether he was acting incident to service.” Id. at 600.

We conclude that Staff Sergeant Re-gan’s duty status was similar to that of Captain Kelly. We do not find it critical that Regan may not have received a pass or leave (only his deposition addresses this point), because under Fort Polk policy he was entitled to off-duty time away from the post without undertaking the formality of the leave and pass paperwork. Regan was in effect on at least a two-day informal pass at the time of the accident. On the duty status continuum, being off post on a pass is sufficiently far from core concerns of Feres as to make the first of the Parker factors weigh in favor of allowing suit.

We next examine whether the location of the injury indicates that Regan’s activity was service oriented.

B. Location of injury

The Toledo Bend Recreation Site, where the boat was rented before being sailed into the Reservoir, is under Fort Polk control and owned by the United States. The district court found a factual dispute about whether the boat was within the geographical boundaries of the MWR site or further out into the public waters of the Reservoir when the injury occurred.

The Government refers us to language in a precedent from another Circuit that the location of the alleged negligence is relevant, not the location of the injury. Bon v. United States, 802 F.2d 1092, 1094 (9th Cir.1986). We make that point because the negligence alleged by Starcraft was in the maintenance and procedures for rental of the boat. That negligence occurred at the MWR facility, which was military property.

Bon does not refer to the place of negligence in explicit contrast to the place of injury. Id. When the injury and the negligence occur in the same place, either word might be used by a court in discussing the facts. Regardless, in the Fifth Circuit, the physical location of the injury is relevant for the incident to service issue, not the location of the possible negligence. Kelly, 26 F.3d at 600 (citing Parker, 611 F.2d at 1014). The purpose of this factor is to determine where the service member was at the time of injury. That fact adds to our understanding beyond just knowing duty status, of whether the service member “was engaged in an activity incident to service.” Parker, 611 F.2d at 1014.

One possible location for the injury here was at the MWR recreation site itself, which has a small cove within its boundaries. It was unclear to the district court and also is to us whether the accident occurred outside that small cove. Regardless of the precise location, there was much less military control there than is exerted at Fort Polk. This factor weighs in favor of suit.

We now examine the final factor, which is whether the activity at the time of the injury served a sufficient military purpose.

C. Activity being performed when injured

The third factor of the Parker test is to “examine the activity being performed at the time of the injury to see if it served some military function.” Parker, 611 F.2d at 1014.

Staff Sergeant Regan was engaged in purely recreational activity. The Army’s investigative report stated that Regan’s activity was “not mission related,” and not related to any tactical or field training. Regan’s presence on the boat was due to his relation with Vandergriff, the renter of the boat. The Government has not suggested any uniquely military function being served by Regan being a guest on the boat. Of course, the purpose of a Morale, Welfare and Recreation facility operated for service members is to improve the morale and welfare of these soldiers, partly through recreation. Such outlets can improve performance of the military mission.

The district court relied on a Ninth Circuit case to hold that the Feres bar applied. Bon, 802 F.2d 1092. In Bon, an active duty sailor’s claim was barred though she was off duty and engaged in recreational activity at the time of her injury. Id. at 1093. The service member in Bon was struck by a motor boat while riding in a canoe rented from a Navy Special Services facility. Use of that facility was restricted to members of the military, Defense Department employees, and their guests and dependents. The commanding officer of the Naval Training Center had direct authority over it. Id. The Ninth Circuit focused on the fact that both the plaintiff and the motorboat driver were active duty service members who were taking part in an activity provided for the benefit of their military service. Id. at 1095. The Bon court concluded that, unlike civilians, the plaintiff and motorboat operator were subject to military discipline for any violations of the Navy’s rules for the rental and operation of the boats, and thus the activity was “incident to service.” Id. at 1095-96.

One of this Circuit’s decisions, also involving a recreational boating accident, distinguished Bon. Kelly, 26 F.3d at 600. Kelly is the precedent concerning an Army captain, while off duty for the weekend in Panama, who rented a catamaran and was killed when the boat’s mast hit low-hanging power lines. Id. at 599. We found Bon inapplicable because there was no evidence that when Captain Kelly was “sailing a catamaran rented from a civilian-run marina,” that he was “directly subject to military control,” while Bon had rented a boat from a Naval recreational facility and was subject to military discipline when using that boat. Id. at 600. The cases may be somewhat closer factually than the distinction suggests. That is because, even though it was not mentioned in the opinion, Captain Kelly had rented his catamaran from a MWR facility. See Brief of Panama Canal Commission, Kelly, 26 F.3d 597 (No. 93-3565), 1993 WL 13101657 at *3-4; cf. Kelly, 26 F.3d at 599 (Court describes the boat-lessor simply as a civilian-run club located at a Naval Station marina). In Bon, the center renting the boat was under the control of the Navy; its purpose was to “provide for Navy personnel and their dependents a varied program of wholesome and constructive off-duty leisure and recreation activities .... ” Bon, 802 F.2d at 1093. Therefore, the Bon and Kelly rental facilities were similar.

Regardless of the precise distinctions between Kelly and Bon, we find little importance to distinguish Kelly and the present case. Both involved service members who rented boats from MWR facilities, subject to the facilities’ regulations. Both involved active duty soldiers who were off duty when the accidents occurred. Both suffered injuries arising from the use of the boats.

A possible way to distinguish this case from Kelly was noted by able counsel for the United States at oral argument, a distinction that creates at least initial doubt about whether Feres was even a valid issue in Kelly. Doubt arises because in Kelly the Panama Canal Commission, not the military, was the allegedly negligent actor for allowing power lines to hang too low over a place in which sailboats were allowed. Kelly, 26 F.3d at 599. Government counsel pointed out at argument that this Commission, perhaps for administrative convenience, fell under the authority of the Secretary of Defense (22 U.S.C. § 3611(a)), a statutory fact which counsel suggested explained why Feres was considered and perhaps also why it was not applied as a bar to suit.

However, there is a more important justification for the Feres analysis in Kelly. That better reason is that the Supreme Court has applied Feres even when the allegedly negligent governmental agency had no connection to the military. In the relevant precedent, a Coast Guardsman’s claims were based upon the alleged negligence of a Federal Aviation Administration air traffic controller. Johnson, 481 U.S. at 681, 107 S.Ct. 2063. This decision was the first time the Supreme Court splintered in its application of the Feres doctrine. Astley, 38 Am. U.L.Rev. at 222. A Coast Guard helicopter pilot, Lieutenant Commander Horton Johnson, was under FAA radar control when responding to a distress call in inclement weather. Johnson, 481 U.S. at 682-83, 107 S.Ct. 2063. The helicopter crashed into the side of a mountain. Johnson’s widow filed a FTCA suit against the United States seeking damages for the injury caused by the alleged negligence by FAA controllers. Id. Even though the FAA was the allegedly negligent actor, and the FAA is not a military agency of the United States, the Feres bar nonetheless applied because Coast Guardsman Johnson’s death was incident to his military duty. Id. at 691, 107 S.Ct. 2063.

That Johnson changed the Feres pattern that a military actor needed to be the negligent agent of the United States is clear when Justice Scalia in dissent stood athwart the road mapped by the majority and said, “no further.”

As the majority acknowledges, however, “all of the cases decided by this Court under Feres have involved allegations of negligence on the part of members of the military.” Ante, at 2066. I would not extend Feres any further. I confess that the line between FTCA suits alleging military negligence and those alleging civilian negligence has nothing to recommend it except that it would limit our clearly wrong decision in Feres and confine the unfairness and irrationality that decision has bred. But that, I think, is justification enough.

Johnson, 481 U.S. at 703, 107 S.Ct. 2063 (Scalia, J., dissenting).

We return to the doctrine’s origins to note that it is “the Government” that is not liable “for injuries to servicemen where the injuries arise out of or are in the course of activity incident to service,” (Feres, 340 U.S. at 146, 71 S.Ct. 153), regardless of what agency of the federal government was the negligent party. By viewing the Panama Canal accident involving Kelly from the perspective of the Supreme Court’s Johnson decision, we find this Court was properly applying controlling precedent. When we found no Feres bar to the suit by the survivors of Army Captain Kelly, we did not silently rely on the fact that the Panama Canal Commission had only an artificial connection to the military. So we look more closely to the reasons for the Kelly decision.

Kelly “was sailing a privately owned catamaran, and no special military rules or regulations applied to govern the conditions of his sailing.” Kelly, 26 F.3d at 600. Allowing an FTCA claim in Kelly would not require a court to second-guess military decisions or impair military discipline. Id. at 601.

Unlike the description of the sailing activity in Kelly (though if Captain Kelly’s boat was rented from an MWR facility, the description may have been inexact), Staff Sergeant Regan was on a boat rented from a military-owned, though civilian-run facility. That facility was far from the post on which Regan served. The renter of the boat, Staff Sergeant Vandergriff, had a few hours before Regan’s arrival agreed to the rules of the rental. Almost certainly, Regan also would have been subject to those rules as a soldier using the facility even though someone else rented the boat.

At this point, it may be useful to sharpen our focus on the purpose of this factor under the Parker test. In determining whether the activity served a military purpose, we agree that the ultimate evaluation is “whether the service-member [stands] in the type of relationship to the military at the time of his or her injury that the occurrences causing the injury arose out of activity incident to military service.” Stephenson v. Stone, 21 F.3d 159, 162 (7th Cir.1994). One commentator summarized her review of the caselaw on whether an activity was related to military service or duties, by stating that the more “attenuated the activity is from the military, the more likely courts will find that the activity was not related to the service member’s military duties.” Brou, 192 Mil. L.Rev. at 28. Admittedly, that conclusion is more a recognition of the obvious than a useful analytical tool, except that the conclusion highlights that courts have been looking at the obvious when making their determinations.

In evaluating whether Regan stood “in the type of relationship to the military at the time of his [ ] injury” that would make “the occurrences causing the injury arose out of activity incident to military service,” we note that his relationship to the Army was coincidental to his injuries and not necessary to them. As a matter of simple causation, Regan was at Toledo Bend, and on this particular boat, because he was a soldier. On the other hand, soldiers’ guests were also allowed to be in exactly the same location doing exactly the same thing. Further, Regan was subject to military penalty for any misconduct at that time, but that would have been the case wherever he was, whatever he was doing. There would have been specific MWR rules applicable to him which could have constituted separate charges and specifications had he violated those. Our determination here of the applicability of the Feres bar reduces to whether largely coincidental factors are enough. When a few, or most, or even all components of a service member’s relationship with the military were unnecessary to the time, location, and manner of the injury, should Feres apply?

To answer that question, we do not venture forth without seeking out the markers already laid. It is inevitable in so fact-specific an area of analysis as is required under Feres, one in which the Supreme Court has stated that there are no “bright-line rules” (Shearer, 473 U.S. at 57, 105 S.Ct. 3039), that the caselaw has some inconsistencies. As much as possible, though, like cases need to be decided alike. The most analogous precedents are those involving recreational activities.

In addressing recreation injuries, some courts barred suit because a service member was receiving a benefit of his military service. For example, the Ninth Circuit determined that no claim could be brought based on the drowning death of a sailor in a rafting trip that had been organized on public waters by a Navy MWR program. Costo v. United States, 248 F.3d 863 (9th Cir.2001). The court stated that its inquiry “begins, and in large measure, ends” with its Bon precedent that we have already discussed. Id. at 867 (citing Bon, 802 F.2d 1092). Because the sailor was on active duty, participating in an activity that was a benefit of his service, and which was conducted by an MWR program under the control of the base commander, Feres applied. Id. One member of the three-member Costo panel dissented, though on the basis that Feres was unconstitutional. Id. at 869 (Ferguson, J., dissenting).

With commendable candor, the majority opinion in Costo noted that its decision went beyond the guideposts set by the Supreme Court:

In citing this litany of cases, it bears note that the Supreme Court has not had occasion to apply Feres nearly so broadly as have the circuit courts. Indeed, it has been suggested that the Supreme Court’s cases could be interpreted to fall within the very limitation-activity based on military duty-that the estates urge us to apply here .... As the estates contend, the Supreme Court cases typically involve “activitfies] incident to service” that implicate military duty, see, e.g., Johnson, 481 U.S. 681, 107 S.Ct. 2063, 95 L.Ed.2d 648, or situations where military discipline was important precisely because it so fundamentally implicated the functioning of the military, see, e.g., Shearer, 473 U.S. at 58, 105 S.Ct. 3039, 87 L.Ed.2d 38. None of these cases involve military-sponsored recreation. But whatever the original scope of the Feres doctrine, it is clear that it has been interpreted throughout the lower courts — and, specifically, by our court — to include military-sponsored recreational programs. Therefore, we are compelled to hold that the estates’ suit is barred.

Costo, 248 F.3d at 869. The Ninth Circuit’s view forms one of the markers as we decide what direction we should take.

At least two precedents prohibited suits brought as a result of incidents involving U.S. Air Force Aero Clubs. Walls v. United States, 832 F.2d 93 (7th Cir.1987); Woodside v. United States, 606 F.2d 134 (6th Cir.1979). These clubs operate on some Air Force bases, are non-appropriated fund instrumentalities of the United States, and promote morale among the active-duty personnel who are eligible for “active membership” in the clubs. Walls, 832 F.2d at 94 n. 2. The clubs are subject to the “strict and pervasive control of the Air Force,” and subject to the overall responsibility of the base commander. Id. The court in Walls relied on the earlier precedent of Woodside, and each found that adverse effects on military discipline would arise from second-guessing the command’s decisions regarding the Aero Clubs. Id. at 95-96.

We find the Aero Club precedents informative, in part because of the distinctions we see. These clubs were supported by and actually located on Air Force bases. The clubs provided an opportunity for airmen “to learn aviation skills under safe, low-cost conditions.” Woodside, 606 F.2d at 142. The Air Force exerts control in part by regulations that minutely detail rights and restrictions. Air Force Instr. 34-217; Air Force Manual 34-232. We agree with finding “this relationship between the Club and the Air Force to be direct and substantial, even though the Club is not an essential or integral part of the military mission of the Air Force.” Woodside, 606 F.2d at 142. The Feres issue of the relationship between an active duty service member and the military examines relative closeness. The connection of Aero Clubs operating on air bases with the Air Force is much closer than an off-post boat rental program with the Army.

The final precedent in the recreational arena that we will consider concerned a suit for injuries sustained at a military horseback riding facility. Hass v. United States, 518 F.2d 1138 (4th Cir.1975). A Marine rented a horse from stables owned and operated by the Marine Corps at its Air Station in Cherry Point, North Carolina. The claim was that the civilian employees of the stables negligently failed to warn of the horse’s propensity to “break its gait and bolt,” causing the Marine injury. Id. at 1139. The Fourth Circuit barred the claim under Feres for these reasons:

We think it might suffice here to bar this suit that (as found by the district court) the stable was owned and operated by the government and that a Marine officer was in charge of it and that servicemen could be disciplined for misconduct while using it. Recreational activity provided by the military can reinforce both morale and health and thus serve the overall military purpose.

Id. at 1141 (footnote omitted). The court focused on several matters, including that the stables were owned and operated by the Marine Corps, and that those who used the stables were subject to discipline for misconduct. As we have pointed out, active duty military members are inevitably subject to discipline regardless of their duty status. An important fact is that it appears the Hass horse stables were on the base itself. Id. at 1139. The Eleventh Circuit interpreted Hass that way, and grouped it analytically with other eases that found Feres applicable when a recreational injury occurred on a military base. Whitley v. United States, 170 F.3d 1061, 1074 & n. 29 (11th Cir.1999).

After this brief caselaw survey, we agree with the observation we earlier quoted from a Ninth Circuit opinion — in the area of military recreational activities, some courts have gone further than the Supreme Court has indicated is necessary. Costo, 248 F.3d at 869. This Circuit has not journeyed as far. We conclude that our decision in Kelly v. Panama Canal Commission, which did not apply the Feres bar, to be a valid, nuanced consideration of the issues that arise in off-post, off-duty recreation. In Kelly, it was not significant that the service member was subject to military discipline, as are all those on active military duty. We found that Captain Kelly was not “directly subject to military control,” a valid short-hand description of a soldier away from his command engaged in recreation that had little if any oversight by the military. In addition, though Kelly had a sailing license issued by the Rodman-Marina Sailing Club, the “civilian-run club located at the Rodman Naval Station” from which the boat was rented, there were “no special rules or regulations applied to govern the conditions of his sailing.” Kelly, 26 F.3d at 599-600. In our case, the boat rental agreement given Staff Sergeant Vander-griff had a minimal checklist of requirements; importantly, the boats were allowed to leave the cove owned by the military and be taken out into the public waters of the Reservoir.

Distinctions can be made, but we find no meaningful differences with Kelly. After reviewing the three Parker factors, we conclude that Staff Sergeant Regan’s injuries were not incident to his service as meant by Feres.

V. Summary and conclusion

The precedents are numerous. The elements of the analysis — the Feres rationales and the factors for an activity being incident to service — are fairly uniform. The nearly-universal point of the caselaw can fairly safely be stated: the further from uniquely military functions an activity may be, and the further from a military base the incident occurs, the less justified is the Feres bar.

This Court has never held that injury suffered by an off-base, off-duty service member when using recreational equipment that is of the kind commonly used by civilians and which was rented by the military to someone else, was incident to service. When a service member is neither on duty nor on a military base and the relationship with the military is largely coincidental and unnecessary to the time, location, and manner of the activity that caused the injury, the argument that an activity is incident to service is at its weakest.

The closest that the issues in Starcraft’s third-party complaint for indemnity will approach the military is that a recreational boat used primarily on non-military waters may not have been properly maintained by the civilian-run Army MWR facility, and that the MWR facility may have failed to follow some of its own procedures during the rental process. Accordingly, some judgments made by members of the military may need to be examined, but that examination would not “involve the judiciary in sensitive military affairs at the expense of military discipline and effectiveness.” Johnson, 481 U.S. at 690-91, 107 S.Ct. 2063. It is true that the Supreme Court has made the incident-to-service analysis controlling for the Feres bar, regardless of whether the particular litigation would intrude alarmingly on military judgments. Stanley, 483 U.S. at 682-83, 107 S.Ct. 3054. However, that the primary rationale for the doctrine is satisfied is at least some confirmation of the validity of the analysis of the incident-to-service issues.

Because of our holding that Staff Sergeant Regan’s activity was not incident to service, Starcraft’s third-party complaint may proceed. The decision of the district court is REVERSED and the cause is REMANDED for proceedings consistent with this opinion.

Notes

. The mission of the Toledo Bend MWR site is to "provide the soldier and authorized patrons the opportunity to participate in recreation activities that are designed to promote relaxation, the development of skills and a sense of well being.”

. We will use the phrase “service member’’ to refer generically to any member of one of the armed services. The rules discussed apply to Marines, Coast Guardsmen, sailors with the Navy, airmen with the Air Force, and, as was Staff Sergeant Regan, soldiers in the Army.

. "Over the years, the Supreme Court has refined and narrowed the foundations of Feres to the point where the military discipline rationale best explains the doctrine.” Note, John Astley, United States v. Johnson: Feres Doctrine Gets New Life and Continues to Grow, 38 Am. U.L.Rev. 185, 217 (1988).

. E.g., Feres, 340 U.S. at 146, 71 S.Ct. 153; Miller v. United States, 42 F.3d 297, 302 (5th Cir.1995); Kelly, 26 F.3d at 600; Parker, 611 F.2d at 1008. Historically, "furlough” was the proper term for authorized absences from duty of American enlisted soldiers. William Winthrop, Military Law And Precedents 562 (G.P.O. reprint 1920) (2d ed. 1896). A U.S. Senate report gave a background on furloughs when the first comprehensive armed forces leave act was adopted after World War II. The reason for the Act was pay equity: by creating a retroactive right to “leave” for enlisted service members to replace a system of purely discretionary furloughs, the Act allowed everyone (not just officers who already had leave rights) coming off active wartime duty to be paid for accumulated leave of as much as 120 days. S.Rep. No. 79-1704, 79th Cong., 2nd Sess., U.S.Code Cong. Service, p. 1342, discussing the Armed Forces Leave Act of August 9, 1946, 60 Stat. 963. The word “furlough” was not used in the 1946 Act, but it did appear in a regulation that was soon superceded. Army Reg. 615-275 (final ed., 5 April 1946). The regulation stated that it was War Department policy to grant to enlisted soldiers an "ordinary furlough” at least once a year, which would be used for "relaxation, recreation, health, and diversion.” Id. ¶ 2. A "furlough” was the term for an authorized absence greater than three days, while "passes” would be for shorter periods. Id. ¶ 1. To implement the 1946 Act, a new regulation was adopted which combined the formerly separate regulations governing leave for officers and authorized absences for enlisted soldiers. Army Reg. 600-115 n.* (20 Aug. 1946). "Furlough” does not appear in the new regulation — regular, lengthy, but discretionary absence from duty was now an obsolete concept. A lengthy authorized absence for enlisted soldiers or for officers was now called “leave.” It would appear that as of the 1946 Act, the word "furlough” was dropped from official Army usage though it lingered as a matter of imprecise jargon. “Leave” and "pass” are now the correct Army terms.

The generous cooperation of a Pentagon librarian, Angela Henson, and one at The Judge Advocate General’s Learning Center and School, Daniel Layering, with a Fifth Circuit librarian, Rosemarie Tominello, to find these dated regulations is appreciated.

. Staff Sergeant Regan's release from active duty was apparently delayed in order that he could continue his medical treatment and re-habilitalion. As of the date of a Government motion filed in August 2006, he was still on active duty.



Concurrence Opinion

DeMOSS, Circuit Judge,

specially concurring:

I concur fully with the majority’s decision to reverse the dismissal of Starcraft’s third-party claim against the United States. I wholeheartedly agree with the majority’s persuasive analysis and ultimate conclusion that the Feres doctrine does not bar the claim.

I write separately to express my personal view that we should have vacated the remand in addition to reversing the dismissal. After the district court erroneously dismissed Starcraft’s third-party claim against the United States, it remanded the case to Louisiana state court pursuant to 28 U.S.C. § 1367(c)(3), which states that a district court may decline to exercise supplemental jurisdiction if “the district court has dismissed all claims over which it has original jurisdiction.” Our decision to reverse the Feres doctrine dismissal automatically renders the section 1367(c)(3) remand erroneous because the dismissal and the remand are inextricably intertwined. For purposes of section 1367(c)(3), “a district court has no discretion to remand a matter in which a federal-law claim still exists.” Burks v. Amerada Hess Corp., 8 F.3d 301, 304 (5th Cir.1993), abrogated on other grounds by Giles v. NYLCare Health Plans, Inc., 172 F.3d 332, 338 (5th Cir.1999).

Starcraft appealed the district court’s judgment dated March 26, 2007, which contained both the dismissal and the remand. Although the notice of appeal gave us jurisdiction to review the remand, Star-craft stated in its briefing that it was not appealing the remand, despite the fact that (1) the validity of the dismissal necessarily implicates the validity of the remand, (2) section 1367(c)(3) remands are reviewable on appeal, and (3) the initial determination of whether the district court has discretion to remand under section 1367(c)(3) is subject to de novo review. Bernhard v. Whitney Nat’l Bank, 523 F.3d 546, 550, No. 07-30464, 2008 WL 879038, at *3 (5th Cir. April 2, 2008); see Hook v. Morrison Milling Co., 38 F.3d 776, 780 (5th Cir.1994) (“If a district court’s decision to remand a case to state court is based on its discretion, then we obviously review that decision for abuse of discretion. The determination of whether the court has that discretion, however, is a legal one, which we review de novo.”) (internal citation omitted). In my view, we have the power to vacate the remand, despite Starcraft’s purported waiver of the issue, in order to give full effect to our judgment reversing the dismissal. I believe the dismissal and the remand are two sides of the same coin and should have been appealed together. See Johnson v. Knorr, 477 F.3d 75, 78, 86-87 (3d Cir.2007) (reversing the dismissal of the federal claim, vacating the section 1367(c)(3) remand of the pendent state law claims, and remanding the case to the district court). Because this case was properly removed and Starcraft has asserted a valid federal claim against the Government, the entire case should be resolved in a single forum: federal district court. Considering the fact that the district court cannot vacate its own remand, New Orleans Pub. Serv., Inc. v. Majoue, 802 F.2d 166, 167 (5th Cir.1986), and because we have the power to dispose of this case “as may be just under the circumstances,” 28 U.S.C. § 2106,1 argue that we should vacate the remand “to obviate further and entirely unnecessary proceedings below.” Grosso v. United States, 390 U.S. 62, 71-72, 88 S.Ct. 709, 19 L.Ed.2d 906 (1968). My colleagues chose to let the parties and the district court resolve the conundrum created by the district court’s precipitous, sua sponte remand of the case to state court and the parties’ failure to address the remand issue on appeal.