Russell Bucklew v. Anne L. Precythe, Director, Missouri Department of Corrections/Opinion of the Court

Notice: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.


No. 17–1184

[April 1, 2019]

Justice Gorsuch delivered the opinion of the Court.

Russell Bucklew concedes that the State of Missouri lawfully convicted him of murder and a variety of other crimes. He acknowledges that the U. S. Constitution permits a sentence of execution for his crimes. He accepts, too, that the State’s lethal injection protocol is constitutional in most applications. But because of his unusual medical condition, he contends the protocol is unconstitutional as applied to him. Mr. Bucklew raised this claim for the first time less than two weeks before his scheduled execution. He received a stay of execution and five years to pursue the argument, but in the end neither the district court nor the Eighth Circuit found it supported by the law or evidence. Now, Mr. Bucklew asks us to overturn those judgments. We can discern no lawful basis for doing so.



In 1996, when Stephanie Ray announced that she wanted to end their relationship, Mr. Bucklew grew violent. He cut her jaw, punched her in the face, and threatened her with a knife. Frightened to remain in the home they had shared, Ms. Ray sought refuge with her children in Michael Sanders’ nearby residence. But then one night Mr. Bucklew invaded that home. Bearing a pistol in each hand, he shot Mr. Sanders in the chest; fired at Mr. Sanders’ 6-year-old son (thankfully, he missed); and pistol-whipped Ms. Ray, this time breaking her jaw. Then Mr. Bucklew handcuffed Ms. Ray, drove her to a secluded spot, and raped her at gunpoint. After a trooper spotted Mr. Bucklew, a shootout followed and he was finally arrested. While all this played out, Mr. Sanders bled to death. As a coda, Mr. Bucklew escaped from jail while awaiting trial and attacked Ms. Ray’s mother with a hammer before he could be recaptured.

After a decade of litigation, Mr. Bucklew was seemingly out of legal options. A jury had convicted him of murder and other crimes and recommended a death sentence, which the court had imposed. His direct appeal had proved unsuccessful. State v. Bucklew, 973 S. W. 2d 83 (Mo. 1998), cert. denied, 525 U. S. 1082 (1999). Separate rounds of state and federal post-conviction proceedings also had failed to yield relief. Bucklew v. State, 38 S. W. 3d 395 (Mo.), cert. denied, 534 U. S. 964 (2001); Bucklew v. Luebbers, 436 F. 3d 1010 (CA8), cert. denied, 549 U. S. 1079 (2006).


As it turned out, though, Mr. Bucklew’s case soon became caught up in a wave of litigation over lethal injection procedures. Like many States, Missouri has periodically sought to improve its administration of the death penalty. Early in the 20th century, the State replaced hanging with the gas chamber. Later in the century, it authorized the use of lethal injection as an alternative to lethal gas. By the time Mr. Bucklew’s post-conviction proceedings ended, Missouri’s protocol called for lethal injections to be carried out using three drugs: sodium thiopental, pancuronium bromide, and potassium chloride. And by that time, too, various inmates were in the process of challenging the constitutionality of the State’s protocol and others like it around the country. See Taylor v. Crawford, 457 F. 3d 902 (CA8 2006); Note, A New Test for Evaluating Eighth Amendment Challenges to Lethal Injections, 120 Harv. L. Rev. 1301, 1304 (2007) (describing flood of lethal injection lawsuits around 2006 that “severely constrained states’ ability to carry out executions”); Denno, The Lethal Injection Quandary: How Medicine Has Dismantled the Death Penalty, 76 Ford. L. Rev. 49, 102–116 (2007).

Ultimately, this Court answered these legal challenges in Baze v. Rees, 553 U. S. 35 (2008). Addressing Kentucky’s similar three-drug protocol, The Chief Justice, joined by Justice Alito and Justice Kennedy, concluded that a State’s refusal to alter its lethal injection protocol could violate the Eighth Amendment only if an inmate first identified a “feasible, readily implemented” alternative procedure that would “significantly reduce a substantial risk of severe pain.” Id., at 52. Justice Thomas, joined by Justice Scalia, thought the protocol passed muster because it was not intended “to add elements of terror, pain, or disgrace to the death penalty.” Id., at 107. Justice Breyer reached the same result because he saw no evidence that the protocol created “a significant risk of unnecessary suffering.” Id., at 113. And though Justice Stevens objected to the continued use of the death penalty, he agreed that petitioners’ evidence was insufficient. Id., at 87. After this Court decided Baze, it denied review in a case seeking to challenge Missouri’s similar lethal injection protocol. Taylor v. Crawford, 487 F. 3d 1072 (2007), cert. denied, 553 U. S. 1004 (2008).

But that still was not the end of it. Next, Mr. Bucklew and other inmates unsuccessfully challenged Missouri’s protocol in state court, alleging that it had been adopted in contravention of Missouri’s Administrative Procedure Act. Middleton v. Missouri Dept. of Corrections, 278 S. W. 3d 193 (Mo.), cert. denied, 556 U. S. 1255 (2009). They also unsuccessfully challenged the protocol in federal court, this time alleging it was pre-empted by various federal statutes. Ringo v. Lombardi, 677 F. 3d 793 (CA8 2012). And Mr. Bucklew sought to intervene in yet another lawsuit alleging that Missouri’s protocol violated the Eighth Amendment because unqualified personnel might botch its administration. That lawsuit failed too. Clemons v. Crawford, 585 F. 3d 1119 (CA8 2009), cert. denied, 561 U. S. 1026 (2010).

While all this played out, pressure from anti-death-penalty advocates induced the company that manufactured sodium thiopental to stop supplying it for use in executions. As a result, the State was unable to proceed with executions until it could change its lethal injection protocol again. This it did in 2012, prescribing the use of a single drug, the sedative propofol. Soon after that, Mr. Bucklew and other inmates sued to invalidate this new protocol as well, alleging that it would produce excruciating pain and violate the Eighth Amendment on its face. After the State revised the protocol in 2013 to use the sedative pentobarbital instead of propofol, the inmates amended their complaint to allege that pentobarbital would likewise violate the Constitution.


Things came to a head in 2014. With its new protocol in place and the necessary drugs now available, the State scheduled Mr. Bucklew’s execution for May 21. But 12 days before the execution Mr. Bucklew filed yet another lawsuit, the one now before us. In this case, he presented an as-applied Eighth Amendment challenge to the State’s new protocol. Whether or not it would cause excruciating pain for all prisoners, as his previous lawsuit alleged, Mr. Bucklew now contended that the State’s protocol would cause him severe pain because of his particular medical condition. Mr. Bucklew suffers from a disease called cavernous hemangioma, which causes vascular tumors—clumps of blood vessels—to grow in his head, neck, and throat. His complaint alleged that this condition could prevent the pentobarbital from circulating properly in his body; that the use of a chemical dye to flush the intravenous line could cause his blood pressure to spike and his tumors to rupture; and that pentobarbital could interact adversely with his other medications.

These latest protocol challenges yielded mixed results. The district court dismissed both the inmates’ facial challenge and Mr. Bucklew’s as-applied challenge. But, at Mr. Bucklew’s request, this Court agreed to stay his execution until the Eighth Circuit could hear his appeal. Bucklew v. Lombardi, 572 U. S. 1131 (2014). Ultimately, the Eighth Circuit affirmed the dismissal of the facial challenge. Zink v. Lombardi, 783 F. 3d 1089 (en banc) (per curiam), cert. denied, 576 U. S. ___ (2015). Then, turning to the as-applied challenge and seeking to apply the test set forth by the Baze plurality, the court held that Mr. Bucklew’s complaint failed as a matter of law to identify an alternative procedure that would significantly reduce the risks he alleged would flow from the State’s lethal injection protocol. Yet, despite this dispositive shortcoming, the court of appeals decided to give Mr. Bucklew another chance to plead his case. The court stressed that, on remand before the district court, Mr. Bucklew had to identify “at the earliest possible time” a feasible, readily implemented alternative procedure that would address those risks. Bucklew v. Lombardi, 783 F. 3d 1120, 1127–1128 (2015) (en banc).

Shortly after the Eighth Circuit issued its judgment, this Court decided Glossip v. Gross, 576 U. S. ___ (2015), rejecting a challenge to Oklahoma’s lethal injection protocol. There, the Court clarified that The Chief Justice’s plurality opinion in Baze was controlling under Marks v. United States, 430 U. S. 188 (1977). In doing so, it reaffirmed that an inmate cannot successfully challenge a method of execution under the Eighth Amendment unless he identifies “an alternative that is ‘feasible, readily implemented, and in fact significantly reduces a substantial risk of severe pain.’” 576 U. S., at ___–___ (slip op., at 12–13). Justice Thomas, joined by Justice Scalia, reiterated his view that the Eighth Amendment “prohibits only those methods of execution that are deliberately designed to inflict pain,” but he joined the Court’s opinion because it correctly explained why petitioners’ claim failed even under the controlling opinion in Baze. Glossip, 576 U. S., at ___ (concurring opinion) (slip op., at 1) (internal quotation marks and alterations omitted).


Despite the Eighth Circuit’s express instructions, when Mr. Bucklew returned to the district court in 2015 he still refused to identify an alternative procedure that would significantly reduce his alleged risk of pain. Instead, he insisted that inmates should have to carry this burden only in facial, not as-applied, challenges. Finally, after the district court gave him “one last opportunity,” App. 30, Mr. Bucklew filed a fourth amended complaint in which he claimed that execution by “lethal gas” was a feasible and available alternative method that would significantly reduce his risk of pain. Id., at 42. Mr. Bucklew later clarified that the lethal gas he had in mind was nitrogen, which neither Missouri nor any other State had ever used to carry out an execution.

The district court allowed Mr. Bucklew “extensive discovery” on his new proposal. 883 F. 3d 1087, 1094 (CA8 2018). But even at the close of discovery in 2017, the district court still found the proposal lacking and granted the State’s motion for summary judgment. By this point in the proceedings, Mr. Bucklew’s contentions about the pain he might suffer had evolved considerably. He no longer complained about circulation of the drug, the use of dye, or adverse drug interactions. Instead, his main claim now was that he would experience pain during the period after the pentobarbital started to take effect but before it rendered him fully unconscious. According to his expert, Dr. Joel Zivot, while in this semiconscious “twilight stage” Mr. Bucklew would be unable to prevent his tumors from obstructing his breathing, which would make him feel like he was suffocating. Dr. Zivot declined to say how long this twilight stage would last. When pressed, however, he referenced a study on euthanasia in horses. He claimed that the horses in the study had displayed some amount of brain activity, as measured with an electroencephalogram (or EEG), for up to four minutes after they were given a large dose of pentobarbital. Based on Dr. Zivot’s testimony, the district court found a triable issue as to whether there was a “substantial risk” that Mr. Bucklew would “experience choking and an inability to breathe for up to four minutes” if he were executed by lethal injection. App. 827. Even so, the court held, Mr. Bucklew’s claim failed because he had produced no evidence that his proposed alternative, execution by nitrogen hypoxia, would significantly reduce that risk.

This time, a panel of the Eighth Circuit affirmed. The panel held that Mr. Bucklew had produced no evidence that the risk of pain he alleged “would be substantially reduced by use of nitrogen hypoxia instead of lethal injection as the method of execution.” 883 F. 3d, at 1096. Judge Colloton dissented, arguing that the evidence raised a triable issue as to whether nitrogen gas would “render Bucklew insensate more quickly than pentobarbital.” Id., at 1099. The full court denied rehearing en banc over a dissent by Judge Kelly, who maintained that, while prisoners pursuing facial challenges to a state execution protocol must plead and prove an alternative method of execution under Baze and Glossip, prisoners like Mr. Bucklew who pursue as-applied challenges should not have to bear that burden. 885 F. 3d 527, 528 (2018).

On the same day Mr. Bucklew was scheduled to be executed, this Court granted him a second stay of execution. 583 U. S. ___ (2018). We then agreed to hear his case to clarify the legal standards that govern an as-applied Eighth Amendment challenge to a State’s method of carrying out a death sentence. 584 U. S. ___ (2018).


We begin with Mr. Bucklew’s suggestion that the test for lethal injection protocol challenges announced in Baze and Glossip should govern only facial challenges, not as-applied challenges like his. In evaluating this argument, we first examine the original and historical understanding of the Eighth Amendment and our precedent in Baze and Glossip. We then address whether, in light of those authorities, it would be appropriate to adopt a different constitutional test for as-applied claims.


The Constitution allows capital punishment. See Glossip, 576 U. S., at ___–___ (slip op., at 2–4); Baze, 553 U. S., at 47. In fact, death was “the standard penalty for all serious crimes” at the time of the founding. S. Banner, The Death Penalty: An American History 23 (2002) (Banner). Nor did the later addition of the Eighth Amendment outlaw the practice. On the contrary—the Fifth Amendment, added to the Constitution at the same time as the Eighth, expressly contemplates that a defendant may be tried for a “capital” crime and “deprived of life” as a penalty, so long as proper procedures are followed. And the First Congress, which proposed both Amendments, made a number of crimes punishable by death. See Act of Apr. 30, 1790, 1 Stat. 112. Of course, that doesn’t mean the American people must continue to use the death penalty. The same Constitution that permits States to authorize capital punishment also allows them to outlaw it. But it does mean that the judiciary bears no license to end a debate reserved for the people and their representatives.

While the Eighth Amendment doesn’t forbid capital punishment, it does speak to how States may carry out that punishment, prohibiting methods that are “cruel and unusual.” What does this term mean? At the time of the framing, English law still formally tolerated certain punishments even though they had largely fallen into disuse—punishments in which “terror, pain, or disgrace [were] superadded” to the penalty of death. 4 W. Blackstone, Commentaries on the Laws of England 370 (1769). These included such “[d]isgusting” practices as dragging the prisoner to the place of execution, disemboweling, quartering, public dissection, and burning alive, all of which Blackstone observed “savor[ed] of torture or cruelty.” Ibid.

Methods of execution like these readily qualified as “cruel and unusual,” as a reader at the time of the Eighth Amendment’s adoption would have understood those words. They were undoubtedly “cruel,” a term often defined to mean “[p]leased with hurting others; inhuman; hard-hearted; void of pity; wanting compassion; savage; barbarous; unrelenting,” 1 S. Johnson, A Dictionary of the English Language (4th ed. 1773), or “[d]isposed to give pain to others, in body or mind; willing or pleased to torment, vex or afflict; inhuman; destitute of pity, compassion or kindness,” 1 N. Webster, An American Dictionary of the English Language (1828). And by the time of the founding, these methods had long fallen out of use and so had become “unusual.” 4 Blackstone, supra, at 370; Banner 76; Baze, 553 U. S., at 97 (Thomas, J., concurring in judgment); see also Stinneford, The Original Meaning of “Unusual:” The Eighth Amendment as a Bar to Cruel Innovation, 102 Nw. U. L. Rev. 1739, 1770–1771, 1814 (2008) (observing that Americans in the late 18th and early 19th centuries described as “unusual” governmental actions that had “fall[en] completely out of usage for a long period of time”).

Contemporary evidence confirms that the people who ratified the Eighth Amendment would have understood it in just this way. Patrick Henry, for one, warned that unless the Constitution was amended to prohibit “cruel and unusual punishments,” Congress would be free to inflict “tortures” and “barbarous” punishments. 3 Debates on the Federal Constitution 447–448 (J. Elliot 2d ed. 1891). Many early commentators likewise described the Eighth Amendment as ruling out “the use of the rack or the stake, or any of those horrid modes of torture devised by human ingenuity for the gratification of fiendish passion.” J. Bayard, A Brief Exposition of the Constitution of the United States 140 (1833); see B. Oliver, The Rights of an American Citizen 186 (1832) (the Eighth Amendment prohibits such “barbarous and cruel punishments” as “[b]reaking on the wheel, flaying alive, rending asunder with horses,… maiming, mutilating and scourging to death”). Justice Story even remarked that he thought the prohibition of cruel and unusual punishments likely “unnecessary” because no “free government” would ever authorize “atrocious” methods of execution like these. 3 J. Story, Commentaries on the Constitution of the United States §1896, p. 750 (1833).

Consistent with the Constitution’s original understanding, this Court in Wilkerson v. Utah, 99 U. S. 130 (1879), permitted an execution by firing squad while observing that the Eighth Amendment forbade the gruesome methods of execution described by Blackstone “and all others in the same line of unnecessary cruelty.” Id., at 135–136. A few years later, the Court upheld a sentence of death by electrocution while observing that, though electrocution was a new mode of punishment and therefore perhaps could be considered “unusual,” it was not “cruel” in the constitutional sense: “[T]he punishment of death is not cruel, within the meaning of that word as used in the Constitution. [Cruelty] implies… something inhuman and barbarous, something more than the mere extinguishment of life.” In re Kemmler, 136 U. S. 436, 447 (1890).

It’s instructive, too, to contrast the modes of execution the Eighth Amendment was understood to forbid with those it was understood to permit. At the time of the Amendment’s adoption, the predominant method of execution in this country was hanging. Glossip, 576 U. S., at ___ (slip op., at 2). While hanging was considered more humane than some of the punishments of the Old World, it was no guarantee of a quick and painless death. “Many and perhaps most hangings were evidently painful for the condemned person because they caused death slowly,” and “[w]hether a hanging was painless or painful seems to have been largely a matter of chance.” Banner 48, 170. The force of the drop could break the neck and sever the spinal cord, making death almost instantaneous. But that was hardly assured given the techniques that prevailed at the time. More often it seems the prisoner would die from loss of blood flow to the brain, which could produce unconsciousness usually within seconds, or suffocation, which could take several minutes. Id., at 46–47; J. Laurence, The History of Capital Punishment 44–46 (1960); Gardner, Executions and Indignities: An Eighth Amendment Assessment of Methods of Inflicting Capital Punishment, 39 Ohio St. L. J. 96, 120 (1978). But while hanging could and often did result in significant pain, its use “was virtually never questioned.” Banner 170. Presumably that was because, in contrast to punishments like burning and disemboweling, hanging wasn’t “intended to be painful” and the risk of pain involved was considered “unfortunate but inevitable.” Ibid.; see also id., at 48.

What does all this tell us about how the Eighth Amendment applies to methods of execution? For one thing, it tells us that the Eighth Amendment does not guarantee a prisoner a painless death—something that, of course, isn’t guaranteed to many people, including most victims of capital crimes. Glossip, 576 U. S., at ___ (slip op., at 4). Instead, what unites the punishments the Eighth Amendment was understood to forbid, and distinguishes them from those it was understood to allow, is that the former were long disused (unusual) forms of punishment that intensified the sentence of death with a (cruel) “‘superadd[ition]’” of “‘terror, pain, or disgrace.’” Baze, 553 U. S., at 48; accord, id., at 96 (Thomas, J., concurring in judgment).

This Court has yet to hold that a State’s method of execution qualifies as cruel and unusual, and perhaps understandably so. Far from seeking to superadd terror, pain, or disgrace to their executions, the States have often sought more nearly the opposite, exactly as Justice Story predicted. Through much of the 19th century, States experimented with technological innovations aimed at making hanging less painful. See Banner 170–177. In the 1880s, following the recommendation of a commission tasked with finding “‘the most humane and practical method known to modern science of carrying into effect the sentence of death,’” the State of New York replaced hanging with electrocution. Glossip, 576 U. S., at ___ (slip op., at 2). Several States followed suit in the “‘“belief that electrocution is less painful and more humane than hanging.”’” Ibid. Other States adopted lethal gas after concluding it was “‘the most humane [method of execution] known to modern science.’” Ibid. And beginning in the 1970s, the search for less painful modes of execution led many States to switch to lethal injection. Id., at ___ (slip op., at 3); Baze, 553 U. S., at 42, 62; see also Banner 178–181, 196–197, 297. Notably, all of these innovations occurred not through this Court’s intervention, but through the initiative of the people and their representatives.

Still, accepting the possibility that a State might try to carry out an execution in an impermissibly cruel and unusual manner, how can a court determine when a State has crossed the line? The Chief Justice’s opinion in Baze, which a majority of the Court held to be controlling in Glossip, supplies critical guidance. It teaches that where (as here) the question in dispute is whether the State’s chosen method of execution cruelly superadds pain to the death sentence, a prisoner must show a feasible and readily implemented alternative method of execution that would significantly reduce a substantial risk of severe pain and that the State has refused to adopt without a legitimate penological reason. See Glossip, 576 U. S., at ___–___ (slip op., at 12–13); Baze, 553 U. S., at 52. Glossip left no doubt that this standard governs “all Eighth Amendment method-of-execution claims.” 576 U. S., at ___ (slip op., at 1).

In reaching this conclusion, Baze and Glossip recognized that the Eighth Amendment “does not demand the avoidance of all risk of pain in carrying out executions.” Baze, 553 U. S., at 47. To the contrary, the Constitution affords a “measure of deference to a State’s choice of execution procedures” and does not authorize courts to serve as “boards of inquiry charged with determining ‘best practices’ for executions.” Id., at 51–52, and nn. 2–3. The Eighth Amendment does not come into play unless the risk of pain associated with the State’s method is “substantial when compared to a known and available alternative.” Glossip, 576 U. S., at ___ (slip op., at 13); see Baze, 553 U. S., at 61. Nor do Baze and Glossip suggest that traditionally accepted methods of execution—such as hanging, the firing squad, electrocution, and lethal injection—are necessarily rendered unconstitutional as soon as an arguably more humane method like lethal injection becomes available. There are, the Court recognized, many legitimate reasons why a State might choose, consistent with the Eighth Amendment, not to adopt a prisoner’s preferred method of execution. See, e. g., Glossip, 576 U. S., at ___–___ (slip op., at 13–14) (a State can’t be faulted for failing to use lethal injection drugs that it’s unable to procure through good-faith efforts); Baze, 553 U. S., at 57 (a State has a legitimate interest in selecting a method it regards as “preserving the dignity of the procedure”); id., at 66 (Alito, J., concurring) (a State isn’t required to modify its protocol in ways that would require the involvement of “persons whose professional ethics rules or traditions impede their participation”).

As we’ve seen, two Members of the Court whose votes were essential to the judgment in Glossip argued that establishing cruelty consistent with the Eighth Amendment’s original meaning demands slightly more than the majority opinion there (or the Baze plurality opinion it followed) suggested. Instead of requiring an inmate to establish that a State has unreasonably refused to alter its method of execution to avoid a risk of unnecessary pain, Justice Thomas and Justice Scalia contended that an inmate must show that the State intended its method to inflict such pain. See Glossip, 576 U. S., at ___ (Thomas, J., concurring) (slip op., at 1); Baze, 553 U. S., at 94–107 (Thomas, J., concurring in judgment). But revisiting that debate isn’t necessary here because, as we’ll see, the State was entitled to summary judgment in this case even under the more forgiving BazeGlossip test. See Part III, infra.


Before turning to the application of Baze and Glossip, however, we must confront Mr. Bucklew’s argument that a different standard entirely should govern as-applied challenges like his. He admits that Baze and Glossip supply the controlling test in facial challenges to a State’s chosen method of execution. But he suggests that he should not have to prove an alternative method of execution in his as-applied challenge because “certain categories” of punishment are “manifestly cruel… without reference to any alternative methods.” Brief for Petitioner 41–42 (internal quotation marks omitted). He points to “‘burning at the stake, crucifixion, [and] breaking on the wheel’” as examples of “categorically” cruel methods. Ibid. And, he says, we should use this case to add to the list of “categorically” cruel methods any method that, as applied to a particular inmate, will pose a “substantial and particular risk of grave suffering” due to the inmate’s “unique medical condition.” Id., at 44.

The first problem with this argument is that it’s foreclosed by precedent. Glossip expressly held that identifying an available alternative is “a requirement of all Eighth Amendment method-of-execution claims” alleging cruel pain. 576 U. S., at ___ (slip op., at 1) (emphasis added). And just as binding as this holding is the reasoning underlying it. Distinguishing between constitutionally permissible and impermissible degrees of pain, Baze and Glossip explained, is a necessarily comparative exercise. To decide whether the State has cruelly “superadded” pain to the punishment of death isn’t something that can be accomplished by examining the State’s proposed method in a vacuum, but only by “compar[ing]” that method with a viable alternative. Glossip, 576 U. S., at ___ (slip op., at 13); see Baze, 553 U. S., at 61. As Mr. Bucklew acknowledges when speaking of facial challenges, this comparison “provides the needed metric” to measure whether the State is lawfully carrying out an execution or inflicting “gratuitous” pain. Brief for Petitioner 42–43. Yet it is that very comparison and needed metric Mr. Bucklew would now have us discard. Nor does he offer some persuasive reason for overturning our precedent. To the contrary, Mr. Bucklew simply repeats the same argument the principal dissent offered and the Court expressly and thoughtfully rejected in Glossip. Just as Mr. Bucklew argues here, the dissent there argued that “certain methods of execution” like “burning at the stake” should be declared “categorically off-limits.” And just as Mr. Bucklew submits here, the dissent there argued that any other “intolerably painful” method of execution should be added to this list. 576 U. S., at ___–___ (Sotomayor, J., dissenting) (slip op., at 23–24). Mr. Bucklew’s submission, thus, amounts to no more than a headlong attack on precedent.

Mr. Bucklew’s argument fails for another independent reason: It is inconsistent with the original and historical understanding of the Eighth Amendment on which Baze and Glossip rest. As we’ve seen, when it comes to determining whether a punishment is unconstitutionally cruel because of the pain involved, the law has always asked whether the punishment “superadds” pain well beyond what’s needed to effectuate a death sentence. And answering that question has always involved a comparison with available alternatives, not some abstract exercise in “categorical” classification. At common law, the ancient and barbaric methods of execution Mr. Bucklew cites were understood to be cruel precisely because—by comparison to other available methods—they went so far beyond what was needed to carry out a death sentence that they could only be explained as reflecting the infliction of pain for pain’s sake. Meanwhile, hanging carried with it an acknowledged and substantial risk of pain but was not considered cruel because that risk was thought—by comparison to other known methods—to involve no more pain than was reasonably necessary to impose a lawful death sentence. See supra, at 9–12.

What does the principal dissent have to say about all this? It acknowledges that Glossip’s comparative requirement helps prevent facial method-of-execution claims from becoming a “backdoor means to abolish” the death penalty. Post, at 8 (opinion of Breyer, J.). But, the dissent assures us, there’s no reason to worry that as-applied method-of-execution challenges might be used that way. This assurance misses the point. As we’ve explained, the alternative-method requirement is compelled by our understanding of the Constitution, not by mere policy concerns.

With that, the dissent is left only to rehash the same argument that Mr. Bucklew offers. The dissent insists that some forms of execution are just categorically cruel. Post, at 10–11. At first and like others who have made this argument, the dissent offers little more than intuition to support its conclusion. Ultimately, though, even it bows to the necessity of something firmer. If a “comparator is needed” to assess whether an execution is cruel, the dissent tells us, we should compare the pain likely to follow from the use of a lethal injection in this case with the pain-free use of lethal injections in mine-run cases. Post, at 10. But that’s just another way of saying executions must always be carried out painlessly because they can be carried out painlessly most of the time, a standard the Constitution has never required and this Court has rejected time and time again. Supra, at 12. To determine whether the State is cruelly superadding pain, our precedents and history require asking whether the State had some other feasible and readily available method to carry out its lawful sentence that would have significantly reduced a substantial risk of pain.

That Mr. Bucklew and the dissent fail to respect the force of our precedents—or to grapple with the understanding of the Constitution on which our precedents rest—is more than enough reason to reject their view that as-applied and facial challenges should be treated differently. But it turns out their position on this score suffers from further problems too—problems that neither Mr. Bucklew nor the dissent even attempts to address.

Take this one. A facial challenge is really just a claim that the law or policy at issue is unconstitutional in all its applications. So classifying a lawsuit as facial or as-applied affects the extent to which the invalidity of the challenged law must be demonstrated and the corresponding “breadth of the remedy,” but it does not speak at all to the substantive rule of law necessary to establish a constitutional violation. Citizens United v. Federal Election Comm’n, 558 U. S. 310, 331 (2010). Surely it would be strange for the same words of the Constitution to bear entirely different meanings depending only on how broad a remedy the plaintiff chooses to seek. See Gross v. United States, 771 F. 3d 10, 14–15 (CADC 2014) (“‘[T]he substantive rule of law is the same for both [facial and as-applied] challenges’”); Brooklyn Legal Servs. Corp. v. Legal Servs. Corp., 462 F. 3d 219, 228 (CA2 2006) (the facial/as-applied distinction affects “the extent to which the invalidity of a statute need be demonstrated,” not “the substantive rule of law to be used”). And surely, too, it must count for something that we have found not a single court decision in over 200 years suggesting that the Eighth Amendment’s meaning shifts in this way. To the contrary, our precedent suggests just the opposite. In the related context of an Eighth Amendment challenge to conditions of confinement, we have seen “no basis whatever” for applying a different legal standard to “deprivations inflicted upon all prisoners” and those “inflicted upon particular prisoners.” Wilson v. Seiter, 501 U. S. 294, 299, n. 1 (1991).

Here’s yet another problem with Mr. Bucklew’s argument: It invites pleading games. The line between facial and as-applied challenges can sometimes prove “amorphous,” Elgin v. Department of Treasury, 567 U. S. 1, 15 (2012), and “not so well defined,” Citizens United, 558 U. S., at 331. Consider an example. Suppose an inmate claims that the State’s lethal injection protocol violates the Eighth Amendment when used to execute anyone with a very common but not quite universal health condition. Should such a claim be regarded as facial or as-applied? In another context, we sidestepped a debate over how to categorize a comparable claim—one that neither sought “to strike [the challenged law] in all its applications” nor was “limited to plaintiff’s particular case”—by concluding that “[t]he label is not what matters.” Doe v. Reed, 561 U. S. 186, 194 (2010). To hold now, for the first time, that choosing a label changes the meaning of the Constitution would only guarantee a good deal of litigation over labels, with lawyers on each side seeking to classify cases to maximize their tactical advantage. Unless increasing the delay and cost involved in carrying out executions is the point of the exercise, it’s hard to see the benefit in placing so much weight on what can be an abstruse exercise.

Finally, the burden Mr. Bucklew must shoulder under the BazeGlossip test can be overstated. An inmate seeking to identify an alternative method of execution is not limited to choosing among those presently authorized by a particular State’s law. Missouri itself seemed to acknowledge as much at oral argument. Tr. of Oral Arg. 65. So, for example, a prisoner may point to a well-established protocol in another State as a potentially viable option. Of course, in a case like that a court would have to inquire into the possibility that one State possessed a legitimate reason for declining to adopt the protocol of another. See supra, at 13–14. And existing state law might be relevant to determining the proper procedural vehicle for the inmate’s claim. See Hill v. McDonough, 547 U. S. 573, 582–583 (2006) (if the relief sought in a 42 U. S. C. §1983 action would “foreclose the State from implementing the [inmate’s] sentence under present law,” then “recharacterizing a complaint as an action for habeas corpus might be proper”). But the Eighth Amendment is the supreme law of the land, and the comparative assessment it requires can’t be controlled by the State’s choice of which methods to authorize in its statutes. In light of this, we see little likelihood that an inmate facing a serious risk of pain will be unable to identify an available alternative—assuming, of course, that the inmate is more interested in avoiding unnecessary pain than in delaying his execution.


Having (re)confirmed that anyone bringing a method of execution claim alleging the infliction of unconstitutionally cruel pain must meet the BazeGlossip test, we can now turn to the question whether Mr. Bucklew is able to satisfy that test. Has he identified a feasible and readily implemented alternative method of execution the State refused to adopt without a legitimate reason, even though it would significantly reduce a substantial risk of severe pain? Because the case comes to us after the entry of summary judgment, this appeal turns on whether Mr. Bucklew has shown a genuine issue of material fact warranting a trial.


We begin with the question of a proposed alternative method. Through much of this case and despite many opportunities, Mr. Bucklew refused to identify any alternative method of execution, choosing instead to stand on his argument that Baze and Glossip’s legal standard doesn’t govern as-applied challenges like his (even after the Eighth Circuit rejected that argument). Only when the district court warned that his continued refusal to abide this Court’s precedents would result in immediate dismissal did Mr. Bucklew finally point to nitrogen hypoxia. The district court then afforded Mr. Bucklew “extensive discovery” to explore the viability of that alternative. 883 F. 3d, at 1094. But even after all that, we conclude Mr. Bucklew has failed for two independent reasons to present a triable question on the viability of nitrogen hypoxia as an alternative to the State’s lethal injection protocol.

First, an inmate must show that his proposed alternative method is not just theoretically “‘feasible’” but also “‘readily implemented.’” Glossip, 576 U. S., at ___–___ (slip op., at 12–13). This means the inmate’s proposal must be sufficiently detailed to permit a finding that the State could carry it out “relatively easily and reasonably quickly.” McGehee v. Hutchinson, 854 F. 3d 488, 493 (CA8 2017); Arthur v. Commissioner, Ala. Dept. of Corrections, 840 F. 3d 1268, 1300 (CA11 2016). Mr. Bucklew’s bare-bones proposal falls well short of that standard. He has presented no evidence on essential questions like how nitrogen gas should be administered (using a gas chamber, a tent, a hood, a mask, or some other delivery device); in what concentration (pure nitrogen or some mixture of gases); how quickly and for how long it should be introduced; or how the State might ensure the safety of the execution team, including protecting them against the risk of gas leaks. Instead of presenting the State with a readily implemented alternative method, Mr. Bucklew (and the principal dissent) point to reports from correctional authorities in other States indicating that additional study is needed to develop a protocol for execution by nitrogen hypoxia. See App. 697 (Oklahoma grand jury report recommending that the State “retain experts” and conduct “further research” to “determine how to carry out the sentence of death by this method”); id., at 736 (report of Louisiana Dept. of Public Safety & Corrections stating that “[r]esearch… is ongoing” to develop a nitrogen hypoxia protocol). That is a proposal for more research, not the readily implemented alternative that Baze and Glossip require.

Second, and relatedly, the State had a “legitimate” reason for declining to switch from its current method of execution as a matter of law. Baze, 553 U. S., at 52. Rather than point to a proven alternative method, Mr. Bucklew sought the adoption of an entirely new method—one that had “never been used to carry out an execution” and had “no track record of successful use.” McGehee, 854 F. 3d, at 493. But choosing not to be the first to experiment with a new method of execution is a legitimate reason to reject it. In Baze we observed that “no other State ha[d] adopted” the one-drug protocol the inmates sought and they had “proffered no study showing” their one-drug protocol would be as effective and humane as the State’s existing three-drug protocol. 553 U. S., at 57. Under those circumstances, we held as a matter of law that Kentucky’s refusal to adopt the inmates’ proffered protocol could not “constitute a violation of the Eighth Amendment.” Ibid. The Eighth Amendment prohibits States from dredging up archaic cruel punishments or perhaps inventing new ones, but it does not compel a State to adopt “untried and untested” (and thus unusual in the constitutional sense) methods of execution. Id., at 41.[1]


Even if a prisoner can carry his burden of showing a readily available alternative, he must still show that it would significantly reduce a substantial risk of severe pain. Glossip, 576 U. S., at ___ (slip op., at 13); Baze, 553 U. S., at 52. A minor reduction in risk is insufficient; the difference must be clear and considerable. Over the course of this litigation, Mr. Bucklew’s explanation why nitrogen hypoxia meets this standard has evolved significantly. But neither of the two theories he has advanced in this Court turns out to be supported by record evidence.

First, Mr. Bucklew points to several risks that he alleges could result from use of the State’s lethal injection protocol that would not be present if the State used nitrogen gas. For example, he says the execution team might try to insert an IV into one of his peripheral veins, which could cause the vein to rupture; or the team might instead use an allegedly painful “cut-down” procedure to access his femoral vein. He also says that he might be forced to lie flat on his back during the execution, which could impair his breathing even before the pentobarbital is administered. And he says the stress from all this could cause his tumors to bleed, further impairing his breathing. These risks, we may assume, would not exist if Mr. Bucklew were executed by his preferred method of nitrogen hypoxia.

The problem with all of these contentions is that they rest on speculation unsupported, if not affirmatively contradicted, by the evidence in this case. Nor does the principal dissent contend otherwise. So, for example, uncontroverted record evidence indicates that the execution team will have discretion to adjust the gurney to whatever position is in Mr. Bucklew’s best medical interests. 883 F. 3d, at 1092, n. 3; App. 531. Moreover, the State agreed in the district court that it would not try to place an IV in Mr. Bucklew’s compromised peripheral veins. Id., at 820; see Brief for Appellant in No. 17–3052 (CA8), p. 7. And, assuming without granting that using a cut-down would raise issues under the Eighth Amendment—but see Nooner v. Norris, 594 F. 3d 592, 604 (CA8 2010) (holding otherwise)—the State’s expert, Dr. Michael Antognini, testified without contradiction that it should be possible to place an IV in Mr. Bucklew’s femoral vein without using a cut-down procedure, App. 350. Mr. Bucklew responds by pointing to the warden’s testimony that he once saw medical staff perform a cut-down as part of an execution; but there’s no evidence that what the warden saw was an attempt to access a femoral vein, as opposed to some other vein.

Moreover, to the extent the record is unclear on any of these issues, Mr. Bucklew had ample opportunity to conduct discovery and develop a factual record concerning exactly what procedures the State planned to use. He failed to do so—presumably because the thrust of his constitutional claim was that any attempt to execute him via lethal injection would be unconstitutional, regardless of the specific procedures the State might use. As the court of appeals explained: “Having taken the position that any lethal injection procedure would violate the Eighth Amendment,” Mr. Bucklew “made no effort to determine what changes, if any, the [State] would make in applying its lethal injection protocol” to him, and he “never urged the district court to establish a suitable fact-finding procedure… to define the as-applied lethal injection protocol [the State] intends to use.” 883 F. 3d, at 1095–1096.[2] Second, Mr. Bucklew contends that the lethal injection itself will expose him to a substantial risk of severe pain that could be eliminated by adopting his preferred method. He claims that once the sedative pentobarbital is injected he will “lose the ability to manage” the tumors in his airway and, as a result, will experience a “sense of suffocation” for some period of time before the State’s sedative renders him fully unconscious. Brief for Petitioner 12–13. “It is during this in-between twilight stage,” according to his expert, Dr. Zivot, “that Mr. Bucklew is likely to experience prolonged feelings of suffocation and excruciating pain.” App. 234. Mr. Bucklew admits that similar feelings of suffocation could occur with nitrogen, the only difference being the potential duration of the so-called “twilight stage.” He contends that with nitrogen the stage would last at most 20 to 30 seconds, while with pentobarbital it could last up to several minutes.

But here again the record contains insufficient evidence to permit Mr. Bucklew to avoid summary judgment. For starters, in the courts below Mr. Bucklew maintained he would have trouble managing his airway only if he were forced to lie supine, which (as we’ve explained) the evidence shows he won’t be. (The dissenters don’t address this point.) But even indulging his new claim that he will have this difficulty regardless of position, he still has failed to present colorable evidence that nitrogen would significantly reduce his risk of pain. We can assume for argument’s sake that Mr. Bucklew is correct that with nitrogen the twilight stage would last 20 to 30 seconds. The critical question, then, is how long that period might last with pentobarbital. The State’s expert, Dr. Antognini, testified that pentobarbital, too, would render Mr. Bucklew fully unconscious and incapable of experiencing pain within 20 to 30 seconds. Id., at 299–301, 432–433. Dr. Zivot disagreed; but when he was asked how long he thought the twilight stage would last with pentobarbital, his testimony was evasive. Eventually, he said his “number would be longer than” 20 to 30 seconds, but he declined to say how much longer. Id., at 195. Instead, he referenced a 2015 study on euthanasia in horses. He said the study found that when horses were given a large dose of pentobarbital (along with other drugs), they exhibited “isoelectric EEG”—a complete absence of detectable brain activity—after 52 to 240 seconds. Id., at 194–196. The district court assumed Dr. Zivot meant that “pain might be felt until measurable brain activity ceases” and that, extrapolating from the horse study, it might take up to four minutes for pentobarbital to “induc[e] a state in which [Mr. Bucklew] could no longer sense that he is choking or unable to breathe.” The district court acknowledged, however, that this might be “a generous interpretation of Dr. Zivot’s testimony.” Id., at 822, and n. 5.

In fact, there’s nothing in the record to suggest that Mr. Bucklew will be capable of experiencing pain for significantly more than 20 to 30 seconds after being injected with pentobarbital. For one thing, Mr. Bucklew’s lawyer now admits that Dr. Zivot “crossed up the numbers” from the horse study. Tr. of Oral Arg. 7–8, 11–12. The study actually reported that the horses displayed isoelectric EEG between 2 and 52 seconds after infusion of pentobarbital was completed, with an average time of less than 24 seconds. App. 267. So if anything, the horse study appears to bolster Dr. Antognini’s time estimate. For another thing, everyone now also seems to acknowledge that isoelectric EEG is the wrong measure. Dr. Zivot never claimed the horses were capable of experiencing pain until they reached isoelectric EEG. And Mr. Bucklew’s lawyer now concedes that doctors perform major surgery on human patients with measurable EEG readings, which strongly suggests that Mr. Bucklew will be insensible to pain before reaching isoelectric EEG. Tr. of Oral Arg. 9. Finally, the record evidence even allows the possibility that nitrogen could increase the risk of pain. Because Dr. Zivot declined to testify about the likely effects of nitrogen gas, Mr. Bucklew must rely on Dr. Antognini’s testimony. And while Dr. Antognini did say he thought nitrogen’s “onset of action” could be “relatively fast,” App. 458, he added that the effects of nitrogen could vary depending on exactly how it would be administered—information Mr. Bucklew hadn’t provided. Indeed, he stated that “depending on… how it’s used, you might get more suffering from nitrogen gas than you would have” from the State’s current protocol. Id., at 460–461.

Of course, the principal dissent maintains that Dr. Zivot’s testimony supports an inference that pentobarbital might cause Mr. Bucklew to suffer for a prolonged period. But its argument rests on a number of mistakes about the record. For example, the dissent points to Dr. Zivot’s remark that, with pentobarbital, “‘the period of time between receiving the injection and death could range over a few minutes to many minutes.’” Post, at 4, 6 (quoting App. 222). From this, the dissent concludes that Mr. Bucklew may suffer for “up to several minutes.” Post, at 1, 6, 9. But everyone agrees that the relevant question isn’t how long it will take for Mr. Bucklew to die, but how long he will be capable of feeling pain. Seeking to address the problem, the dissent next points to another part of Dr. Zivot’s testimony and says it means Mr. Bucklew could experience pain during the entire time between injection and death. Post, at 6, 13 (quoting App. 222). But the dissent clips the relevant quotation. As the full quotation makes clear, Dr. Zivot claimed that Mr. Bucklew might be unable to “maintain the integrity of his airway” until he died—but he carefully avoided claiming that Mr. Bucklew would be capable of feeling pain until he died.[3] To avoid this problem, the dissent quotes Dr. Zivot’s assertions that pentobarbital might not produce “‘rapid unconsciousness’” and that Mr. Bucklew’s suffering with pentobarbital could be “‘prolonged.’” Post, at 4–6, 13 (quoting App. 233–234). But Dr. Zivot’s statements here, too, fail to specify how long Mr. Bucklew is likely to be able to feel pain. The hard fact is that, when Dr. Zivot was finally compelled to offer a view on this question, his only response was to refer to the horse study. Id., at 195–196. The dissent’s effort to suggest that Dr. Zivot “did not rely exclusively or even heavily on that study,” post, at 7, is belied by (among other things) Mr. Bucklew’s own brief in this Court, which asserted that the twilight stage during which he might feel pain could last “between 52 and 240 seconds,” based entirely on a citation of Dr. Zivot’s incorrect testimony about the horse study. Brief for Petitioner 13.

In sum, even if execution by nitrogen hypoxia were a feasible and readily implemented alternative to the State’s chosen method, Mr. Bucklew has still failed to present any evidence suggesting that it would significantly reduce his risk of pain. For that reason as well, the State was entitled to summary judgment on Mr. Bucklew’s Eighth Amendment claim.[4]


“Both the State and the victims of crime have an important interest in the timely enforcement of a sentence.” Hill, 547 U. S., at 584. Those interests have been frustrated in this case. Mr. Bucklew committed his crimes more than two decades ago. He exhausted his appeal and separate state and federal habeas challenges more than a decade ago. Yet since then he has managed to secure delay through lawsuit after lawsuit. He filed his current challenge just days before his scheduled execution. That suit has now carried on for five years and yielded two appeals to the Eighth Circuit, two 11th-hour stays of execution, and plenary consideration in this Court. And despite all this, his suit in the end amounts to little more than an attack on settled precedent, lacking enough evidence even to survive summary judgment—and on not just one but many essential legal elements set forth in our case law and required by the Constitution’s original meaning.

The people of Missouri, the surviving victims of Mr. Bucklew’s crimes, and others like them deserve better. Even the principal dissent acknowledges that “the long delays that now typically occur between the time an offender is sentenced to death and his execution” are “excessive.” Post, at 16. The answer is not, as the dissent incongruously suggests, to reward those who interpose delay with a decree ending capital punishment by judicial fiat. Post, at 18. Under our Constitution, the question of capital punishment belongs to the people and their representatives, not the courts, to resolve. The proper role of courts is to ensure that method-of-execution challenges to lawfully issued sentences are resolved fairly and expeditiously. Courts should police carefully against attempts to use such challenges as tools to interpose unjustified delay. Last-minute stays should be the extreme exception, not the norm, and “the last-minute nature of an application” that “could have been brought” earlier, or “an applicant’s attempt at manipulation,” “may be grounds for denial of a stay.” Hill, 547 U. S., at 584 (internal quotation marks omitted). So, for example, we have vacated a stay entered by a lower court as an abuse of discretion where the inmate waited to bring an available claim until just 10 days before his scheduled execution for a murder he had committed 24 years earlier. See Dunn v. Ray, 586 U. S. ___ (2019).[5] If litigation is allowed to proceed, federal courts “can and should” protect settled state judgments from “undue interference” by invoking their “equitable powers” to dismiss or curtail suits that are pursued in a “dilatory” fashion or based on “speculative” theories. Id., at 584–585.


The judgment of the court of appeals is



  1. While this case has been pending, a few States have authorized nitrogen hypoxia as a method of execution. See 2018 Ala. Acts no. 2018–353 (allowing condemned inmates to elect execution by nitrogen hypoxia); 2017 Miss. Laws ch. 406, p. 905 (authorizing execution by nitrogen hypoxia only if lethal injection is held unconstitutional or is otherwise unavailable); 2015 Okla. Sess. Laws ch. 75, p. 244 (same). In March 2018, officials in Oklahoma announced that, due to the unavailability of lethal injection drugs, the State would use nitrogen gas for its executions going forward. See Williams, Oklahoma Proposes To Use Nitrogen Gas for Executions by Asphyxiation, N. Y. Times, Mar. 15, 2018, p. A22. But Oklahoma has so far been unable to find a manufacturer willing to sell it a gas delivery device for use in executions. See Clay, State Not Ready for Executions, The Oklahoman, Jan. 27, 2019, p. A1. To date, no one in this case has pointed us to an execution in this country using nitrogen gas.
  2. While the district court allowed discovery on many other matters, Mr. Bucklew protests that it did not permit him to learn the identities of the lethal injection execution team members, to depose them, or to inquire into their qualifications, training, and experience. Like the Eighth Circuit, we see no abuse of discretion in the district court’s discovery rulings. As the district court explained, Mr. Bucklew argues that there is no way he may be constitutionally executed by lethal injection, even with modifications to the State’s lethal injection protocol. And in a case like that, discovery into such granular matters as who administers the protocol simply is not relevant.
  3. Here’s the full quotation, with the portion quoted by the dissent underlined:

    “As a result of his inability to maintain the integrity of his airway for the period of time beginning with the injection of the Pentobarbital solution and ending with Mr. Bucklew’s death several minutes to as long as many minutes later, Mr. Bucklew would be highly likely to experience feelings of ‘air hunger’ and the excruciating pain of prolonged suffocation resulting from the complete obstruction of his airway by the large vascular tumor.” App. 222.

  4. The State contends that Mr. Bucklew’s claim should fail for yet another reason: because, in the State’s view, the evidence does not show that he is very likely to suffer “‘severe pain’” cognizable under the Eighth Amendment. Glossip v. Gross, 576 U. S. ___, ___ (2015) (slip op., at 13) (quoting Baze v. Rees, 553 U. S. 35, 52 (2008); emphasis added). We have no need, however, to address that argument because (as explained above) Mr. Bucklew fails even to show that a feasible and readily available alternative could significantly reduce the pain he alleges.
  5. Seeking to relitigate Dunn v. Ray, the principal dissent asserts that that case involved no undue delay because the inmate “brought his claim only five days after he was notified” that the State would not allow his spiritual adviser to be present with him in the execution chamber itself, although it would allow the adviser to be present on the other side of a glass partition. Post, at 17. But a state statute listed “[t]he spiritual adviser of the condemned” as one of numerous individuals who would be allowed to “be present at an execution,” many of whom—such as “newspaper reporters,” “relatives or friends of the condemned person,” and “the victim’s immediate family members”—obviously would not be allowed into the chamber itself. Ala. Code §15–18–83 (2018). The inmate thus had long been on notice that there was a question whether his adviser would be allowed into the chamber or required to remain on the other side of the glass. Yet although he had been on death row since 1999, and the State had set a date for his execution on November 6, 2018, he waited until January 23, 2019—just 15 days before the execution—to ask for clarification. He then brought a claim 10 days before the execution and sought an indefinite stay. This delay implicated the “strong equitable presumption” that no stay should be granted “where a claim could have been brought at such a time as to allow consideration of the merits without requiring entry of a stay.” Hill v. McDonough, 547 U. S. 573, 584 (2006).