dissenting:
The war has ended and the troops are home, but to the majority this case lives on. I respectfully dissent.
*1384Federal jurisdiction extends only to cases or controversies, and the case or controversy must exist at every stage of a lawsuit’s pendency in federal court. Lewis v. Continental Bank Corp., 494 U.S. 472, 110 S.Ct. 1249, 1253-54, 108 L.Ed.2d 400 (1990); Clarke v. United States, 915 F.2d 699, 700-01 (D.C.Cir.1990) (en banc). This requirement, like standing and ripeness and other doctrines in the article III family, is particular and individual: Both the plaintiff and the defendant must maintain a personal stake in the suit’s outcome. Lewis, 110 S.Ct. at 1254; Clarke, 915 F.2d at 701. The Does’ lawsuit grew directly and solely out of the events leading up to, and culminating in, the allied nations’ recent victory over Iraq in Operation Desert Storm. Now that the war has ended, the Does’ dispute with the defendants is purely hypothetical, and their interest in the outcome, academic.
As the majority explains, the Department of Defense, preparing for combat, last fall reassessed its ability to prevent and treat the physiological hazards associated with chemical and biological warfare. The Department recognized that two drugs would help allied soldiers survive if Iraq engaged in such warfare. Neither drug had been approved by the FDA, however, and then-current FDA regulations (codified at 21 C.F.R. § 50.23 (1990) and known as rule 23) permitted waiver of the statutory informed-consent requirement only under certain conditions, including the condition that the person taking the drug actually be incapable of giving consent. The Department of Defense reaffirmed its commitment to the informed-consent requirement but explained that war plans had convinced it that “another circumstance should be recognized ... in which it would be consistent with the statute and ethically appropriate ... to ‘deem it not feasible’ to obtain informed consent of the patient.” Informed Consent for Human Drugs and Biologies, 55 Fed.Reg. 52,814, 52,815 (1990) (citation omitted).
On December 21, 1990, less than one month before Desert Shield became Desert Storm, the FDA added subsection (d) to rule 23. The newly-added subsection allows the Department of Defense to request a determination from the Commissioner of the FDA that obtaining informed consent is not feasible, but only, as the majority notes, under certain strictly limited circumstances. See id. at 52,817.
First, there must be “a specific military operation involving combat or the immediate threat of combat.” Second, the Department must support in writing the conclusions of Department doctors and investigators that there exist “special military combat ... circumstances in which, in order to facilitate the accomplishment of the military mission, preservation of the health of the individual and the safety of other personnel require that a particular treatment be provided to a specified group of military personnel.” Third, the Department must determine that the treatment should be provided “without regard to what might be any individual’s personal preference for no treatment or for some alternative treatment.” Fourth, the Department must confirm that an institutional review board has approved the waiver request. Fifth, the Commissioner of the FDA must determine that there exists “no available satisfactory alternative therapy.” Sixth, the Commissioner must determine that “withholding treatment would be contrary to the best interests of military personnel.” In reaching that decision, the Commissioner must take into account “all pertinent factors,” including the safety of the investigational drug, its effectiveness, the context in which it will be administered, the nature of the disease or condition to be treated, and the information to be provided to those taking the drug on the benefits and risks of taking or not taking the drug. If the Commissioner does decide to grant the Department’s request, the waiver automatically expires in a year or when the Department tells the Commissioner that the military threat has ended, whichever comes first.
John Doe, stationed at the time in Saudi Arabia, filed this suit (with his wife, whose cause of action is derivative) on January 11, 1991, fewer than five days before Desert Storm began. Seeking a declaration that rule 23(d) is invalid and an injunction fore-*1385ing Secretary Sullivan to revoke it, the Does challenged rule 23(d) under three different legal theories. Each of the challenges alleged that rule 23(d), on its face and as-applied, is inconsistent with a statute or the Constitution.
Each of the challenges is moot. On February 27, President Bush halted offensive military operations in the Gulf, and on March 15, the Assistant Secretary of Defense for Health Affairs, pursuant to rule 23(d), informed the Commissioner of the FDA that “the specific military operations creating the need for the use of these two drug products without informed consent has ended.” The rule is no longer in play. Doe thus will not be subjected to the rule, and he retains no live, personal stake in the outcome of the suit that challenges it.
The Does concede, and the majority acknowledges, that their as-applied challenges are moot. Summoning up what the majority calls the “exception to the mootness doctrine for controversies that are ‘capable of repetition, yet evading review,’ ” ante at 1375-76, however, the Does continue to press their facial challenges to the rule, and the majority proceeds to adjudicate them. I find this case moot for the reasons that follow. Because I believe that we have no power to decide this lawsuit, I express no view of the merits.
The majority correctly restates the two requirements of what the Supreme Court recently emphasized remains “the exceptional situation,” Lewis, 110 S.Ct. at 1255: “(1) the challenged action [must be] in its duration too short to be fully litigated prior to its cessation or expiration, and (2) there [must be] a reasonable expectation that the same complaining party [will] be subjected to the same action yet again.” Ante at 1376 (quoting Weinstein v. Bradford, 423 U.S. 147, 149, 96 S.Ct. 347, 348, 46 L.Ed.2d 350 (1975)). This exceptional case, the Court has stressed, requires that there be a reasonable likelihood that the complaining party personally be subjected to the challenged action. When the party asserting jurisdiction proves that the challenged action meets these conditions, the case is not moot. In my view, the Does have failed to meet the second of these conditions: I see little expectation, much less a reasonable one, that John Doe will ever be subjected to the operation of rule 23(d) again.
For the action that the Does challenge to recur, a sequence of events must take place, each of those events mandated by rule 23(d) itself. Yet the occurrence of any of those events is improbable alone and virtually unimaginable in sequence. First, the United States would have to embark on “a specific military operation involving combat or the immediate threat of combat.” Second, the military operation would have to involve medical hazards that require treatment with an investigational drug. Third, the Department of Defense would have to determine that special military circumstances made informed consent infeasible. Fourth, an institutional review board would have to agree. Fifth, the Commissioner of the FDA would have to determine that no alternative to the investi-gational drug exists. Sixth, the Commissioner would have to evaluate the Department’s request in light of the factors set out in rule 23(d) and then agree to it. Most important, John Doe himself would have to be involved in the war and subject personally to the medical threat. If the Does fail to show that there exists a “reasonable expectation” that any one of these conditions will occur, this court may not exercise jurisdiction over their claims.
Minimizing most of these contingencies, the majority concentrates on citing evidence meant to show that there exists a risk of chemical warfare. See ante at 1377; cf. Whiteside, Annals of the Cold War — The Yellow-Rain Complex (pts. 1 & 2), The New Yorker (Feb. 11, 1991 & Feb. 18, 1991). Chemical warfare has indeed been waged, with tragic results, on unarmed civilians in Kurdistan, but the majority surely overstates those risks to American soldiers, such as Doe — after all, American soldiers have not been the victims of organized chemical attack since the First World War. Moreover, some of the evidence that the majority cites suggests that the likelihood of chemical warfare will diminish in the future: President Bush recently announced our country’s commit*1386ment to the multilateral Chemical Weapons Convention, currently being negotiated, in which, as the President explained,
we are formally forswearing the use of chemical weapons for any reason, including retaliation, against any state, ... and will propose that all states follow suit. Further, the United States unconditionally commits itself to the destruction of all our stocks of chemical weapons within 10 years ... and will propose that all other states do likewise.
27 Weekly Comp. Pres. Doc. 599, 600 (May 13, 1991).
In any event, it is not enough that a war involving chemical weapons may occur at some time in the future. Nor is it enough that that war might lead the Department of Defense to invoke the strict conditions of rule 23(d) or that each of the conditions might be met, possibilities almost as speculative as the possibility that American troops might face a chemical attack. Instead, there must be a reasonable likelihood that John Doe personally will be involved in that war, that he personally will face the medical threat that requires the administration of investigational drugs, and that he personally will be subject to the operation of rule 23(d). The majority focuses on rule 23(d) in the abstract — and in the process forgets about Doe, the plaintiff.
Yet it was Doe, of course, who by alleging concrete injury made this case a “case” instead of a request for an advisory opinion. The majority concludes that Doe “has not lost a ‘personal stake’ in this case” because “Doe, though among the troops brought safely home, remains a soldier.” Ante at 1378. But Doe’s remaining a soldier only lets the majority bring Doe’s being subjected to rule 23(d) from the realm of the impossible (since the rule applies only to soldiers) to the realm of the extraordinarily unlikely. This is all the majority can do — for the record is virtually silent. Is Doe about to be discharged, this year, or next? Does he serve in the infantry, or behind a desk? Has he been assigned for the rest of his tour to permanent duty in the United States? If sent back overseas, will Doe serve in England or Germany, or in the Middle East? Doe tells us nothing, and without more information we cannot determine whether Doe himself faces any likelihood of being subjected to rule 23(d).
Doe, in his silence, has failed to carry his burden of proving the existence of federal jurisdiction. All that we do know is that the war with Iraq ended in February and that on April 29, Doe left the Gulf and returned to the United States with his unit. We also know that Doe, as far as the record reveals, had never before been subjected to combat giving rise to the need for investigational drugs, and in the thirty years since the Food, Drug & Cosmetic Act has required informed consent before the drugs can be administered, the government had never before this fall asked that the requirement be waived. We noted in Clarke that when a court “estimates] the likelihood of an event’s occurring in the future, a natural starting point is how often it has occurred in the past.” 915 F.2d at 704; see ante at 1376. This is the sole, isolated occasion. I see no reasonable possibility of it happening to John Doe in the future.
In my view, this case is moot. I would therefore vacate the decision of the district court and remand this case with instructions to dismiss.