Blueford v. Arkansas,
566 U.S. 599 (2012)

Annotate this Case
  • Syllabus  | 
  • Opinion (John G. Roberts, Jr.)  | 
  • Dissent (Sonia Sotomayor)



No. 10–1320



on writ of certiorari to the supreme court of arkansas

[May 24, 2012]

     Justice Sotomayor, with whom Justice Ginsburg and Justice Kagan join, dissenting.

     The Double Jeopardy Clause “unequivocally prohibits a second trial following an acquittal.” Arizona v. Washington, 434 U. S. 497, 503 (1978) . To implement this rule, our cases have articulated two principles. First, an acquittal occurs if a jury’s decision, “whatever its label, actually rep-resents a resolution, correct or not, of some or all of the factual elements of the offense charged.” United States v. Martin Linen Supply Co., 430 U. S. 564, 571 (1977) . Sec-ond, a trial judge may not defeat a defendant’s entitlement to “the verdict of a tribunal he might believe to be favorably disposed to his fate” by declaring a mistrial before deliberations end, absent a defendant’s consent or a “ ‘manifest necessity’ ” to do so. United States v. Jorn, 400 U. S. 470, 486, 481 (1971) (plurality opinion) (quoting United States v. Perez, 9 Wheat. 579, 580 (1824)).

     Today’s decision misapplies these longstanding prin-ciples. The Court holds that petitioner Alex Blueford was not acquitted of capital or first-degree murder, even though the forewoman of the Arkansas jury empaneled to try him announced in open court that the jury was “unanimous against” convicting Blueford of those crimes. Nor, the Court concludes, did the Double Jeopardy Clause oblige the trial judge to take any action to give effect to the jury’s unambiguous decision before declaring a mistrial as to those offenses. The Court thus grants the State what the Constitution withholds: “the proverbial ‘second bite at the apple.’ ” Burks v. United States, 437 U. S. 1, 17 (1978) .

     I respectfully dissent.



     The bar on retrials following acquittals is “the most fundamental rule in the history of double jeopardy jurisprudence.” Martin Linen, 430 U. S., at 571; see, e.g., United States v. Ball, 163 U. S. 662, 671 (1896) ; 4 W. Blackstone, Commentaries on the Laws of England 329 (1769). This prohibition stops the State, “with all its resources and power,” from mounting abusive, harassing reprosecutions, Green v. United States, 355 U. S. 184, 187 (1957) , which subject a defendant to “embarrassment, expense, anxiety, and insecurity, and the possibility that he may be found guilty even though innocent,” United States v. DiFrancesco, 449 U. S. 117, 136 (1980) .

     In ascertaining whether an acquittal has occurred, “form is not to be exalted over substance.” Sanabria v. United States, 437 U. S. 54, 66 (1978) . Rather, we ask whether the factfinder has made “a substantive determination that the prosecution has failed to carry its burden.” Smith v. Massachusetts, 543 U. S. 462, 468 (2005) . Ju-risdictions have different procedures respecting the announcement of verdicts and the entry of judgments, but that diversity has no constitutional significance. Jeopardy terminates upon a determination, however characterized, that the “evidence is insufficient” to prove a defendant’s “factual guilt.” Smalis v. Pennsylvania, 476 U. S. 140, 144 (1986) . Thus, we have treated as acquittals a trial judge’s directed verdict of not guilty, Smith, 543 U. S., at 468; an appellate reversal of a conviction for insufficiency of the evidence, Burks, 437 U. S., at 10; and, most pertinent here, a jury’s announcement of a not guilty verdict that was “not followed by any judgment,” Ball, 163 U. S., at 671.

     A straightforward application of that principle suffices to decide this case. Arkansas is a classic “acquittal-first” or “hard-transition” jurisdiction. See generally People v. Richardson, 184 P. 3d 755, 764, n. 7 (Colo. 2008). Arkansas’ model jury instructions require a jury to complete its deliberations on a greater offense before it may consider a lesser. 1 Ark. Model Jury Instr., Crim., No. 302 (2d ed. 1994). As a matter of Arkansas law, “[b]efore it may consider any lesser-included offense, the jury must first determine that the proof is insufficient to convict on the greater offense. Thus, the jury must, in essence, acquit the defendant of the greater offense before considering his or her guilt on the lesser-included offense.” Hughes v. State, 347 Ark. 696, 706–707, 66 S. W. 3d 645, 651 (2002). [ 1 ]

     Here, the trial judge instructed Blueford’s jury to consider the offenses in order, from the charged offense of capital murder to the lesser included offenses of first-degree murder, manslaughter, and negligent homicide. The judge told the jury to proceed past capital murder only upon a unanimous finding of a “reasonable doubt” as to that offense—that is, upon an acquittal. See In re Winship, 397 U. S. 358, 363 (1970) . The State’s closing arguments repeated this directive: “[B]efore you can consider a lesser included of capital murder, you must first, all 12, vote that this man is not guilty of capital murder.” App. 55. And the forewoman’s colloquy with the judge leaves no doubt that the jury understood the instructions to mandate unanimous acquittal on a greater offense as a prerequisite to consideration of a lesser: The forewoman reported that the jury had not voted on negligent homicide because the jurors “couldn’t get past the manslaughter” count on which they were deadlocked. Id., at 65.

     In this context, the forewoman’s announcement in open court that the jury was “unanimous against” conviction on capital and first-degree murder, id., at 64–65, was an acquittal for double jeopardy purposes. [ 2 ] Per Arkansas law, the jury’s determination of reasonable doubt as to those offenses was an acquittal “in essence.” Hughes, 347 Ark., at 707, 66 S. W. 3d, at 651. By deciding that the State “had failed to come forward with sufficient proof,” the jury resolved the charges of capital and first-degree murder adversely to the State. Burks, 437 U. S., at 10. That acquittal cannot be reconsidered without putting Blueford twice in jeopardy.

     Green and Price v. Georgia, 398 U. S. 323 (1970) , bolster that conclusion. In Green, the jury convicted the defendant on the lesser included offense of second-degree murder without returning a verdict on the charged offense of first-degree murder. This Court concluded that this result was an “implicit acquittal” on the greater offense of first-degree murder, barring retrial. 355 U. S., at 190. The defendant “was forced to run the gantlet once on that charge and the jury refused to convict him. When given the choice between finding him guilty of either first or second degree murder it chose the latter.” Ibid.; see also Price, 398 U. S., at 329 (“[T]his Court has consistently refused to rule that jeopardy for an offense continues after an acquittal, whether that acquittal is express or implied by a conviction on a lesser included offense when the jury was given a full opportunity to return a verdict on the greater charge” (footnote omitted)). Notably, Green acknowledged that its finding of an “implicit acquittal” was an “assumption,” because the jury had made no express statement with respect to the greater offense. 355 U. S., at 190–191.

     Blueford’s position is even stronger because his jury was not silent on the murder counts, but announced that it was “unanimous against” conviction. And the trial judge specifically instructed the jury to consider manslaughter only after acquitting Blueford of the murder counts. Courts in several acquittal-first jurisdictions have held that a jury’s deadlock on a lesser included offense justifies the assumption that the jury acquitted on any greater offenses. See State v. Tate, 256 Conn. 262, 283–285, 773 A. 2d. 308, 323–324 (2001); Stone v. Superior Ct. of San Diego Cty., 31 Cal. 3d 503, 511–512, n. 5, 646 P. 2d 809, 815, n. 5 (1982). That assumption is not even necessary here because the jury unmistakably announced acquittal.


     The majority holds that the forewoman’s announcement was not an acquittal because it “was not a final resolution of anything.” Ante, at 6. In the majority’s view, the jury might have revisited its decisions on the murder counts during the 31 minutes of deliberations that followed the forewoman’s announcement. We cannot know whether the jury did so, the majority reasons, because the jury was discharged without confirming that it remained “unanimous against” convicting Blueford of capital and first-degree murder. Ante, at 6–8. [ 3 ]

     Putting to one side the lack of record evidence to support this speculation—by far the more plausible inference is that the jurors spent those 31 minutes attempting to resolve their deadlock on manslaughter—I do not agree that the jury was free to reconsider its decisions when its deliberations resumed. “A verdict of acquittal on the issue of guilt or innocence is, of course, absolutely final.” Bullington v. Missouri, 451 U. S. 430, 445 (1981) . The jury heard instructions and argument that it was required unanimously to acquit on capital and first-degree murder before it could reach manslaughter. And as the forewoman’s colloquy makes plain, the jury followed those instructions scrupulously. There is no reason to believe that the jury’s vote was anything other than a verdict in substance—that is, a “final collective decision . . . reached after full deliberation, consideration, and compromise among the individual jurors.” Harrison v. Gillespie, 640 F. 3d 888, 906 (CA9 2011) (en banc). And when that decision was announced in open court, it became entitled to full double jeopardy protection. See, e.g., Commonwealth v. Roth, 437 Mass. 777, 796, 776 N. E. 2d 437, 450–451 (2002) (declining to give effect to “ ‘ “the verdict received from the lips of the foreman in open court” ’ ” would “elevate form over substance”); Stone, 31 Cal. 3d, at 511, 646 P. 2d, at 814–815 (“[I]n determining what verdict, if any, a jury intended to return, the oral declaration of the jurors endorsing the result is the true return of the verdict” (internal quotation marks omitted)); see also, e.g., Dixon v. State, 29 Ark. 165, 171 (1874) (technical defect in verdict “is of no consequence whatever, for the verdict need not be in writing, but may be announced by the foreman of the jury orally”); State v. Mills, 19 Ark. 476 (1858) (“The verdict was of no validity until delivered, by the jury, in Court”).

     The majority’s example of a jury that takes a preliminary vote on greater offenses, advances to the consideration of a lesser, and then returns to a greater, is inapposite. See ante, at 7. In the majority’s example, the jury has not announced its vote in open court. Moreover, the instructions in this case did not contemplate that the jury’s deliberations could take the course that the majority imagines. Arkansas’ model instruction requires acquittal as a prerequisite to consideration of a lesser offense, and the Double Jeopardy Clause entitles an acquittal to final-ity. Indeed, the purpose of an acquittal-first instruction is to ensure careful and conclusive deliberation on a greater offense. See United States v. Tsanas, 572 F. 2d 340, 346 (CA2 1978) (Friendly, J.) (acquittal-first instruction avoids “the danger that the jury will not adequately discharge its duties with respect to the greater offense, and instead will move too quickly to the lesser”). True, Arkansas’ instruction does not expressly forbid reconsideration, but it does not expressly permit reconsideration either. In any event, nothing indicates that the jury’s announced decisions were tentative, compromises, or mere steps en route to a final verdict, and the Double Jeopardy Clause demands that ambiguity be resolved in favor of the defendant. See Downum v. United States, 372 U. S. 734, 738 (1963) .

     The fact that the jury was not given the express option of acquitting on individual offenses is irrelevant. See ante, at 3, 9. Arkansas law ascribes no significance to the presence of such options on a verdict form. See Rowland v. State, 263 Ark. 77, 85, 562 S. W. 2d 590, 594 (1978) (“The jury may prepare and present its own form of verdict”). The lack of a state procedural vehicle for the entry of a judgment of acquittal does not prevent the recognition of an acquittal for constitutional purposes. See Hudson v. Louisiana, 450 U. S. 40, 41, n. 1 (1981) .

     Finally, the majority’s distinction of Green and Price is unavailing. The majority observes that Green and Price, unlike this case, involved final decisions. Ante, at 8. As I have explained, I view the forewoman’s announcements of acquittal in this case as similarly final. In any event, Green clarified that the defendant’s “claim of former jeopardy” was “not based on his previous conviction for second degree murder but instead on the original jury’s refusal to convict him of first degree murder.” 355 U. S., at 190, n. 11; accord, id., at 194, n. 14. That is, the jury’s silence on the greater offense spoke with sufficient clarity to justify the assumption of acquittal and to invoke the Double Jeopardy Clause. Id., at 191; see also Price, 398 U. S., at 329. In light of the forewoman’s announcement, this is an a fortiori case.

     In short, the Double Jeopardy Clause demands an inquiry into the substance of the jury’s actions. Blueford’s jury had the option to convict him of capital and first-degree murder, but expressly declined to do so. That ought to be the end of the matter.



     Even if the majority were correct that the jury might have reconsidered an acquitted count—a doubtful assumption for the reasons just explained—that would not defeat Blueford’s double jeopardy claim. It “has been long established as an integral part of double jeopardy jurisprudence” that “a defendant could be put in jeopardy even in a prosecution that did not culminate in a conviction or an acquittal.” Crist v. Bretz, 437 U. S. 28, 34 (1978) . This rule evolved in response to the “abhorrent” practice under the Stuart monarchs of terminating prosecutions, and thereby evading the bar on retrials, when it appeared that the Crown’s proof might be insufficient. Washington, 434 U. S., at 507; see, e.g., Ireland’s Case, 7 How. St. Tr. 79, 120 (1678). Accordingly, retrial is barred if a jury is discharged before returning a verdict unless the defendant consents or there is a “manifest necessity” for the discharge. Perez, 9 Wheat., at 580; see also King v. Perkins, 90 Eng. Rep. 1122 (K. B. 1698).

     In Perez, this Court explained that “manifest necessity” is a high bar: “[T]he power ought to be used with the greatest caution, under urgent circumstances, and for very plain and obvious causes.” 9 Wheat., at 580. Since Perez, this Court has not relaxed the showing required. See, e.g., Washington, 434 U. S., at 506 (requiring a “ ‘high degree’ ” of necessity); Downum, 372 U. S., at 736 (“imperious necessity”); see also, e.g., United States v. Coolidge, 25 F. Cas. 622, 623 (No. 14,858) (CC Mass. 1815) (Story, J.) (“extraordinary and striking circumstances”). Before de-claring a mistrial, therefore, a trial judge must weigh heavily a “defendant’s valued right to have his trial completed by a particular tribunal.” Wade v. Hunter, 336 U. S. 684, 689 (1949) . And in light of the historical abuses against which the Double Jeopardy Clause guards, a trial judge must tread with special care where a mistrial would “help the prosecution, at a trial in which its case is going badly, by affording it another, more favorable opportunity to convict the accused.” Gori v. United States, 367 U. S. 364, 369 (1961) ; see Green, 355 U. S., at 188.

     A jury’s genuine inability to reach a verdict constitutes manifest necessity. But in an acquittal-first jurisdiction, a jury that advances to the consideration of a lesser included offense has not demonstrated an inability to decide a de-fendant’s guilt or innocence on a greater—it has acquit- ted on the greater. Under Green, that is unquestionably true if the jury convicts on the lesser. See id., at 189. It would be anomalous if the Double Jeopardy Clause offered less protection to a defendant whose jury has deadlocked on the lesser and thus convicted of nothing at all. See Stone, 31 Cal. 3d, at 511–512, n. 5, 646 P. 2d, at 815, n. 5.

     I would therefore hold that the Double Jeopardy Clause requires a trial judge, in an acquittal-first jurisdiction, to honor a defendant’s request for a partial verdict before declaring a mistrial on the ground of jury deadlock. Courts in acquittal-first jurisdictions have so held. See, e.g., Tate, 256 Conn., at 285–287, 773 A. 2d, at 324–325; Whiteaker v. State, 808 P. 2d 270, 274 (Alaska App. 1991); Stone, 31 Cal. 3d, at 519, 646 P. 2d, at 820; State v. Pugliese, 120 N. H. 728, 730, 422 A. 2d 1319, 1321 (1980) (per curiam); State v. Castrillo, 90 N. M. 608, 611, 566 P. 2d 1146, 1149 (1977); see also N. Y. Crim. Proc. Law Ann. §310.70 (West 2002). Requiring a partial verdict in an acquittal-first jurisdiction ensures that the jurisdiction takes the bitter with the sweet. In general, an acquittal-first instruction increases the likelihood of conviction on a greater offense. See People v. Boettcher, 69 N. Y. 2d 174, 182, 505 N. E. 2d 594, 597 (1987). True, such an instruction may also result in deadlock on a greater, preventing a State “from obtaining a conviction on the lesser charge that would otherwise have been forthcoming and thus require the expense of a retrial.” Tsanas, 572 F. 2d, at 346. But a State willing to incur that expense loses nothing by overcharging in an acquittal-first regime. At worst, the State enjoys a second opportunity to convict, “with the possibility that the earlier ‘trial run’ will strengthen the prosecution’s case.” Crist, 437 U. S., at 52 (Powell, J., dissenting). If a State wants the benefits of requiring a jury to acquit before compromising, it should not be permitted to deprive a defendant of the corresponding benefits of having been acquitted. The Double Jeopardy Clause expressly prohibits that outcome.

     The majority observes that we “have never required a trial court, before declaring a mistrial because of a hung jury, to consider any particular means of breaking the impasse—let alone to consider giving the jury new options for a verdict.” Ante, at 10 (citing Renico v. Lett, 559 U. S. ___, ___ (2010) (slip op., at 8)). That hands-off approach dilutes Perez beyond recognition. This Court has never excused a trial judge from exercising “scrupulous” care before discharging a jury. Jorn, 400 U. S., at 485 (plural-ity opinion). Rather, we have insisted that a trial judge may not act “irrationally,” “irresponsibly,” or “precipi-tately.” Washington, 434 U. S., at 514–515. Nor have we retreated from the rule that “reviewing courts have an obligation to ensure themselves that . . . the trial judge exercised ‘sound discretion’ in declaring a mistrial.” Id., at 514 (quoting Perez, 9 Wheat., at 580). [ 4 ]


     Even if the Double Jeopardy Clause did not compel that broader rule, the facts of this case confirm that there was no necessity, let alone manifest necessity, for a mistrial. There was no reason for the judge not to have asked the jury, prior to discharge, whether it remained “unanimous against” conviction on capital and first-degree murder. There would have been no intrusion on the jury’s deliberative process. The judge was not required to issue new instructions or verdict forms, allow new arguments, direct further deliberations, or take any other action that might have threatened to coerce the jury. Merely repeating his earlier question would have sufficed. Because the judge failed to take even this modest step—or indeed, to explore any alternatives to a mistrial, or even to make an on-the-record finding of manifest necessity—I conclude that there was an abuse of discretion. See, e.g., id., at 486; see also Washington, 434 U. S., at 525 (Marshall, J., dissenting) (manifest necessity requires showing “that there were no meaningful and practical alternatives to a mistrial, or that the trial court scrupulously considered available alter-natives and found all wanting but a termination of the proceedings”).

     Indeed, the only reason I can divine for the judge’s failure to take this modest step is his misperception of Arkansas law with respect to the transitional instruction. After the colloquy with the forewoman, the judge commented at sidebar that the jurors “haven’t even taken a vote on [negligent homicide]. . . . I don’t think they’ve completed their deliberation. . . . I mean, under any reasonable circumstances, they would at least take a vote on negligent homicide.” App. 65–66. And after the jury retired for the last half-hour of deliberations, the judge said, “I don’t think they have an understanding of really that they don’t have to get past every charge unanimously before they can move to the next charge.” Id., at 69. That misstated Arkansas law as well as the judge’s own instructions. The jury was required to reach a unanimous decision on a greater offense before considering a lesser. See supra, at 2–3. In discharging the jury, the judge said, “Madam Foreman, there seems to be a lot of confusion on the part . . . of the jury about some of the instructions. And because of the confusion and because of the timeliness and the amount of hours that has gone by without being able to reach a verdict, the Court is going to declare a mistrial.” App. 69–70.

     If, as these comments suggest, the judge wrongly believed that the jury was not required to reach unanimity on a greater offense before considering a lesser, then he accorded insufficient finality and weight to the forewoman’s earlier announcement of acquittal on capital and first-degree murder. That mistake of law negates the deference due the judge’s decision to declare a mistrial. The judge explained that the jury was being discharged in part based on its “confusion” with respect to the instructions, when in fact, the confusion was the judge’s. Ibid.; see, e.g., Washington, 434 U. S., at 510, n. 28 (“If the record reveals that the trial judge has failed to exercise the ‘sound discretion’ entrusted to him, the reason for . . . deference by an appellate court disappears”); Illinois v. Somerville, 410 U. S. 458, 469 (1973) (critiquing “erratic” mistrial inquiry); Gori, 367 U. S., at 371, n. 3 (Douglas, J., dissenting) (noting that “[i]n state cases, a second prosecution has been barred where the jury was discharged through the trial judge’s misconstruction of the law,” and collecting cases). And a trial court “by definition abuses its discretion when it makes an error of law.” Koon v. United States, 518 U. S. 81, 100 (1996) .

*  *  *

     At its core, the Double Jeopardy Clause reflects the wisdom of the founding generation, familiar to “ ‘every person acquainted with the history of governments,’ ” that “ ‘state trials have been employed as a formidable engine in the hands of a dominant administration. . . . To prevent this mischief the ancient common law . . . provided that one acquittal or conviction should satisfy the law.’ ” Ex parte Lange, 18 Wall. 163, 171 (1874) (quoting Commonwealth v. Olds, 15 Ky. 137, 139 (1824)). The Double Jeopardy Clause was enacted “ ‘[t]o perpetuate this wise rule, so favorable and necessary to the liberty of the citizen in a government like ours.’ ” 18 Wall., at 171. This case demonstrates that the threat to individual freedom from reprosecutions that favor States and unfairly rescue them from weak cases has not waned with time. Only this Court’s vigilance has.

     I respectfully dissent.


1  The State has taken the same position. See Brief for Appellee in Boyd v. State, No. CR 06–973 (Ark.), p. 13 (“[U]nanimity is the essence of a jury verdict as it pertains to acquitting a defendant of the charged offense and the subsequent consideration of lesser-included offenses”).
2  The jury’s acquittals on the murder counts were unsurprising in light of the deficiencies in the State’s case. For example, Dr. Adam Craig—the medical examiner who autopsied the victim, Matthew McFadden, Jr., and whose testimony was essential to the State’s theory of the crime—was not board certified in anatomical pathology, hav-ing failed the certification examination five times. Dr. Craig took only 2 slides of Matthew’s brain, not the 10 to 20 called for by prevail-ing professional standards. He dismissed Blueford’s explanation for Matthew’s death—that Blueford accidentally knocked Matthew to the floor—on the basis of an outdated paper on child head injuries, acknowledging that he was only “vaguely aware” of a more recent, seminal paper that supported Blueford’s account. Record 390; see Goldsmith & Plunkett, A Biomechanical Analysis of the Causes of Traumatic Brain Injury in Infants and Children, 25 Am. J. Forensic Med. & Pathology 89 (2004). Blueford’s expert pathologist, Dr. Robert Bux, testified that Dr. Craig’s autopsy was inadequate to establish whether Matthew’s death was accidental or intentional. And Blue-ford’s expert pediatrician, Dr. John Galaznik, testified that the State’s theory—that Blueford slammed Matthew into a mattress on the floor—was “not a likely cause” of the boy’s injuries when assessed in view of current medical literature. Record 766. Even the trial judge observed that the State’s proof was “circumstantial at best,” and that this was “probably . . . a lesser included offense case.” Id., at 610.
3  This Court granted certiorari to decide “[w]hether, if a jury deadlocks on a lesser-included offense, the Double Jeopardy Clause bars reprosecution of a greater offense after a jury announces that it has voted against guilt on the greater offense.” Pet. for Cert. i. The major-ity resolves the question presented by determining that the forewoman’s announcements were not final, such that Blueford’s jury did not necessarily deadlock on the lesser included offense of manslaughter. See ante, at 8–9, n. 2. In light of that determination, I do not read today’s opinion to express any view with respect to the requirements of the Double Jeopardy Clause where a jury does deadlock on a lesser in-cluded offense. Cf., e.g., State v. Tate, 256 Conn. 262, 284–285, 773 A. 2d 308, 324–325 (2001); Whiteaker v. State, 808 P. 2d 270, 274 (Alaska App. 1991).
4  The majority’s reliance on Renico, a habeas corpus case decided under the deferential standard of review prescribed by the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), is perplexing. As Renico made clear, the question there was “not whether the trial judge should have declared a mistrial. It is not even whether it was an abuse of discretion for her to have done so—the applicable standard on direct review. The question under AEDPA is instead whether the determination of the Michigan Supreme Court that there was no abuse of discretion was ‘an unreasonable application of . . . clearly established Federal law.’ ” 559 U. S., at ___ (slip op., at 5) (quoting ); accord, 559 U. S., at ___, n. 3 (slip op., at 11, n. 3). Renico thus has little to say about a trial judge’s responsibilities, or this Court’s, on direct review. Cf. Cullen v. Pinholster, 563 U. S. ___, ___ (2011) (slip op., at 30–31).
Disclaimer: Official Supreme Court case law is only found in the print version of the United States Reports. Justia case law is provided for general informational purposes only, and may not reflect current legal developments, verdicts or settlements. We make no warranties or guarantees about the accuracy, completeness, or adequacy of the information contained on this site or information linked to from this site. Please check official sources.