Darius Moss appeals from the district court's1 denial of his initial motion pursuant to 28 U.S.C. § 2255 to set aside his sentence. Moss argues his 360-month sentence for drug law violations was imposed in violation of the rule announced in Apprendi v. New Jersey, 120 S. Ct. 2348 (2000), because drug quantity was not charged in his indictment or submitted to the jury during trial. Because we conclude Moss is foreclosed from collaterally attacking his sentence based on Apprendi, we affirm the judgment of the district court.

Moss was convicted in September 1996 of one count of conspiracy to possess with intent to distribute crack cocaine and one count of possession with intent to distribute crack cocaine, in violation of 21 U.S.C. §§ 841(a)(1) and 846. At Moss's sentencing hearing in July 1997, the district court found by a preponderance of the evidence that Moss was responsible for 1,644.3 grams of crack cocaine, which supported a combined base offense level of 38. The district court added two levels for obstruction of justice, see USSG § 3C1.1 (1995), and two levels for recklessly creating a substantial risk of death or serious bodily injury to another in the course of fleeing from a law enforcement officer, see id. § 3C1.2. Moss's combined adjusted offense level of 42 and a criminal history category III resulted in a sentencing range of 360 months to life. The district court sentenced Moss at the bottom end of the range, imposing concurrent terms of 360 months on the conspiracy count and 240 months on the distribution count.

Moss's conviction and sentence was affirmed on direct appeal, see United States v. Moss, 138 F.3d 742 (8th Cir. 1998), and Moss then filed the present § 2255 motion, which the district court denied. This court subsequently granted Moss a certificate of appealability on the issue of whether Jones v. United States, 526 U.S. 227 (1999), applies to 21 U.S.C. § 841. The Supreme Court held in Jones that serious bodily injury under the federal car-jacking statute, see 18 U.S.C. § 2119(2), is an element of the offense, not a sentencing factor, which must be charged in an indictment and submitted to the jury. Jones, 526 U.S. at 251-52.

Moss's opening brief focuses on the validity of the district court's two-level enhancement for reckless endangerment during flight. He argues that after Jones the government was required to charge reckless endangerment in the indictment and prove to the jury beyond a reasonable doubt that he created a substantial risk of death or injury. Shortly after the opening brief was filed, the Supreme Court issued its decision in Apprendi, in which it held that any fact (other than a prior conviction) which increases the penalty for a crime beyond the maximum statutory penalty authorized by a legislature must be submitted to a jury and proved beyond a reasonable doubt. 120 S. Ct. at 2362-63. Our circuit subsequently held in the context of § 841's quantity-dependent sentencing scheme that Apprendi prohibits the government from seeking to impose a sentence in excess of § 841(b)(1)(C)'s 20-year maximum sentence unless drug quantity is both alleged in the indictment and found beyond a reasonable doubt by a jury. See United States v. Aguayo-Delgado, 220 F.3d 926, 933-34 (8th Cir. 2000).

Following Apprendi and Aguayo-Delgado, Moss now raises the issue of whether his sentence is improper because the district court's drug quantity finding increased his sentence beyond § 841(b)(1)(C)'s 20-year maximum sentence. Although the certificate of appealability was issued prior to Apprendi, that decision is a natural outgrowth of, and closely related to, the Jones issue on which the certificate was granted. We therefore believe we have the authority to decide the Apprendi question raised, and neither party suggests otherwise.

We find no merit to Moss's initial argument that the district court's imposition of the reckless endangerment during flight enhancement is constitutionally unsound after Jones or Apprendi. Moss contends the district court's finding that he recklessly created a substantial risk of death and serious bodily injury increased his sentence beyond § 841(b)(1)(C)'s 20-year maximum sentence. His argument, however, confuses the Guidelines enhancement with § 841(b)(1)(C)'s statutory enhancement when "death or serious bodily injury results from the use" of a controlled substance, which exposes a defendant to a maximum statutory penalty of life imprisonment. The district court's finding related solely to whether Moss's relevant conduct, his flight from law enforcement officers, was a sufficient basis to enhance his Guideline sentence and played no part in exposing Moss to the higher statutory sentencing range. A district court may always find relevant conduct under the Guidelines by a preponderance of the evidence because the Guidelines themselves prohibit a sentence in excess of the statutory maximum sentence authorized for the offense of conviction. See USSG §§ 5G1.1, 5G1.2 (2000); see also United States v. Jones, 248 F.3d 671, 677 (7th Cir.2001) (rejecting argument that relevant conduct must be proven to jury beyond a reasonable doubt).

Moss is correct, however, in his assertion that the district court's drug quantity finding increased his sentence beyond the 20-year maximum, thereby resulting in a violation of the rule announced in Apprendi. The government concedes the constitutional violation but argues Moss is not entitled to relief because (1) Apprendi is a new rule of constitutional law inapplicable to cases on collateral review, see Teague v. Lane, 489 U.S. 288 (1989); and (2) Moss procedurally defaulted the claim by failing to raise it in his direct appeal.2

In Teague, the Supreme Court held that new constitutional rules of criminal procedure cannot be applied retroactively to cases on collateral review unless they fall within an exception to the general rule. 489 U.S. at 311. The Court recognized two such exceptions. Relevant to our inquiry is the exception permitting watershed rules, ones which "implicate the fundamental fairness of the trial," to be raised collaterally.3 Id. at 312 (internal quotations omitted). In Rodgers v. United States, 229 F.3d 704 (8th Cir. 2000) (per curiam), we held that § 2255 forecloses Apprendi claims in a second or successive § 2255 motion because the Supreme Court has not "made" Apprendi retroactive to cases on collateral review. Id. at 706 (discussing the language of § 2255). We subsequently noted in United States v. Nicholson, 231 F.3d 445, 454 n.1 (8th Cir. 2000), that whether an Apprendi challenge raised in an initial § 2255 motion is Teague-barred is an open question in this Circuit. Consistent with the Ninth and Fourth Circuits, and the overwhelming majority of district courts,4 we hold today that Apprendi is not of watershed magnitude and that Teague bars petitioners from raising Apprendi claims on collateral review.

The Supreme Court's Teague inquiry is implicated because Apprendi is obviously a "new rule" subject to the general rule of non-retroactivity. A "new rule" is one that "breaks new ground or imposes a new obligation on the States or the Federal Government. . . . To put it differently, a case announces a new rule if the result was not dictated by precedent existing at the time the defendant's conviction became final." Teague, 489 U.S. at 301. We believe that under either definition, Apprendi announces a new rule.

Prior to Apprendi, every federal circuit to have considered the question had held that drug quantity was a sentencing factor rather than an element of the crime defined in 21 U.S.C. § 841. See United States v. Thomas, 204 F.3d 381, 383 (2d Cir. 2000), cert. granted, judgment vacated, and remanded, ___U.S.___, 121 S. Ct. 749 (2001). This precedent was reaffirmed after Jones was decided. Our own case of United States v. Grimaldo, 214 F.3d 967 (8th Cir. 2000), decided just three weeks before Apprendi, is a case on point. Apprendi unmistakably altered the legal landscape and is easily categorized as a new rule.

The second step in our analysis is to ascertain whether the new constitutional principle announced in Apprendi is a watershed rule of criminal procedure, defined as a rule which implicates both the accuracy and fundamental fairness of criminal proceedings. Teague, 489 U.S. at 312. The Supreme Court has described this exception as encompassing only a "small core of rules requiring observance of those procedures that . . . are implicit in the concept of ordered liberty." O'Dell v. Netherland, 521 U.S. 151, 157 (1997) (quoting Graham v. Collins, 506 U.S. 461, 478 (1993)).5 According to the Court, the "sweeping rule" announced in Gideon v. Wainwright, 372 U.S. 335 (1963), that counsel shall be provided in all criminal trials for serious offenses, is the prototypical example of a watershed ruling. See O'Dell, 521 U.S. at 167; Gray v. Netherland, 518 U.S. 152, 170 (1996). Gideon, according to the Court, announced a rule that contains the "primacy and centrality" necessary to place it within Teague's watershed exception. Saffle v. Parks, 494 U.S. 484, 495 (1990). Apprendi does not fall within the same vein as Gideon's pronouncement that one who is unable to afford a lawyer "cannot be assured a fair trial unless counsel is provided for him." 372 U.S. at 344 (emphasis added). In other words, we do not believe Apprendi's rule recharacterizing certain facts as offense elements that were previously thought to be sentencing factors resides anywhere near that central core of fundamental rules that are absolutely necessary to insure a fair trial.

One might conclude at first blush that Apprendi improves the accuracy of the fact-finding process, the first element of a watershed rule, because it increases the prosecution's burden to establish the factual issues which in turn drive the length of a defendant's sentence. To the extent the decision prevents the government from seeking a punishment in excess of one authorized by the jury's verdict, we would agree the accuracy of a criminal proceeding is improved. As Justice O'Connor explained in Teague, however, the accuracy element within the watershed exception derives from one of the principal functions of habeas corpus, which is to "assure that no man has been incarcerated under a procedure which creates an impermissibly large risk that the innocent will be convicted." Teague, 489 U.S. at 312. The exception therefore applies only to "those new procedures without which the likelihood of an accurate conviction is seriously diminished." Id. at 313 (emphasis added). In this light, it seems arguable whether Apprendi increases the reliability of the guilt-innocence determination at all because the rule does not protect the innocent from conviction, it instead limits the sentencing exposure of those who have been validly convicted.6 To illustrate, we have yet to reverse a conviction for Apprendi error, nor can we fathom a situation where a conviction would be overturned because of an Apprendi error. See, e.g., United States v. Ray, 250 F.3d 596, 603 (8th Cir.2001) (rejecting argument that a new trial is the appropriate remedy for an Apprendi error). We have instead granted relief from the sentence imposed following conviction. It also seems arguable whether the integrity of pre-Apprendi criminal convictions were "seriously" compromised by permitting sentences to be set based upon factors found by a judge under the preponderance standard rather than by a jury under the reasonable doubt standard.

We find it unnecessary to delve further into whether Apprendi increases the accuracy of the trial because a new rule must do more than just improve accuracy, worthy as that goal may be. To fall within the exception, the rule must impart a fundamental procedural right that, like Gideon, is a necessary component of a fair trial. See Sawyer v. Smith, 497 U.S. 227, 242 (1990). "It is . . . not enough under Teague to say that a new rule is aimed at improving the accuracy of trial. More is required. A rule that qualifies under this exception must not only improve accuracy, but also alter our understanding of the bedrock procedural elements essential to the fairness of a proceeding." Id. (internal quotations omitted). One need only peruse the cases, and the "new rules" therein, in which the Supreme Court has rejected the watershed exception's applicability to appreciate how absolutely fundamental the right must be to satisfy the exception. See, e.g., United States v. Mandanici, 205 F.3d 519, 529 (2d Cir.) (describing eleven cases since Teague where the Supreme Court has addressed new rules or proposed new rules and refused to apply those rules retroactively under the watershed exception), cert. denied, ___U.S.___, 121 S. Ct. 190 (2000). Apprendi appears no more "important" to a fair trial than rules previously addressed by the Court, including the rule announced in Batson v. Kentucky, 176 U.S. 79 (1986), which the Court refused to apply retroactively in Teague.

Permitting a judge-found fact to affect the sentence imposed after a valid conviction, even if it is found under a more lenient standard, cannot be said to have resulted in a fundamentally unfair criminal proceeding. As the Fifth Circuit has noted, "one can easily envision a system of 'ordered liberty' in which certain elements of a crime can or must be proved to a judge, not to the jury," United States v. Shunk, 113 F.3d 31, 37 (5th Cir. 1997), and it is not as though defendants have been foreclosed prior to Apprendi from challenging facts that were previously thought to be sentencing considerations. For instance, in Moss's case, a sentencing hearing was held at which Moss had the opportunity to both challenge the government's drug-quantity evidence and present his own evidence relevant to the quantity determination.7

What strikes us as particularly indicative that Apprendi is not on a par with Gideon is that the new rule announced floats and flows with the tide of legislative pronouncements. Similar to the scenario Justice O'Connor presents in her Apprendi dissent, see 120 S. Ct. at 2390, Congress could tomorrow eliminate § 841's quantity-dependent sentencing scheme by imposing a maximum penalty of life imprisonment for any violation of the statute, thereby permitting the Sentencing Guidelines' drug quantity assessments made by a judge using a preponderance standard to inform the judge's actual sentencing decision. That would be constitutionally permissible as the law stands after Apprendi, regardless of whether any quantity was alleged in the indictment. (One has a tendency to forget that the Sentencing Guidelines have passed muster with the Congress.) Yet a defendant convicted and sentenced under today's version of § 841 would be entitled to attack retroactively his conviction, while one convicted and sentenced under tomorrow's version of § 841, even though both defendants were subjected to identical criminal proceedings and received identical sentences, would have no constitutional basis to challenge his sentence. Cf. Sanders, 247 F.3d at 150 (noting that Apprendi's holding that a judge's finding can increase a defendant's sentence within a statutory range "undercuts the argument that it states a bedrock principle as envisioned by Teague"). The fact that a legislative body's determination of the statutory maximum associated with a particular crime determines whether a sentence may be based upon a judge-found fact is incompatible with a right that is absolutely necessary to a fair trial.

Our holding that the rule is not of watershed magnitude is consistent with and supported by our court's previous recognition that an Apprendi violation is not a structural error requiring per se reversal. See United States v. Anderson, 236 F.3d 427, 429 (8th Cir. 2001).8 A structural error "deprive[s] defendants of basic protections without which a criminal trial cannot reliably serve its function as a vehicle for determination of guilt or innocence . . . and no criminal punishment may be regarded as fundamentally fair." Neder v. United States, 527 U.S. 1, 8-9 (1999) (internal quotations omitted). The Supreme Court has had no occasion to contrast a watershed rule with structural error, but we agree with the statement in Sanders that "finding something to be a structural error would seem to be a necessary predicate for a new rule to apply retroactively under Teague." 247 F.3d at 150-51; cf. Shunk, 113 F.3d at 37 (rejecting the proposition that all structural errors fall within the watershed exception). The watershed exception is a habeas principle and carries with it the Supreme Court's precept that final convictions should be preserved. Structural error, in contrast, is essentially a doctrine relevant to direct review (it permits a defendant to raise an argument on appeal that was not raised during the trial) where the same finality concerns do not exist. It is thus logical that a watershed rule must be more "fundamental" than a structural error. It is for these reasons we conclude that Apprendi does not fall within Teague's exception for watershed rules.

Even assuming an Apprendi challenge is not Teague barred, we nonetheless conclude that Moss cannot challenge his sentence on Apprendi grounds because he failed to raise the argument in his direct appeal.9 Because habeas relief is an extraordinary remedy which "will not be allowed to do service for an appeal," significant barriers exist in the path of a petitioner who seeks to raise an argument collaterally which he failed to raise on direct review. See Bousley v. United States, 523 U.S. 614, 621 (1998) (internal citations omitted). More specifically, a claim unraised on direct appeal is procedurally defaulted unless a petitioner can demonstrate (1) cause for the default and actual prejudice or (2) actual innocence. Id. at 622.

Moss contends that cause exists to excuse his default because an Apprendi claim falls within the category of those "novel" claims which justifiably may be raised for the first time in a collateral proceeding.10 The Supreme Court recognized in Bousley that "a claim that 'is so novel that its legal basis is not reasonably available to counsel' may constitute cause for a procedural default." Id. at 622 (emphasis added) (quoting Reed v. Ross, 468 U.S. 1, 16 (1984)). We recognize the Apprendi decision caused an about-face in our understanding of what constitutes an element of an offense, but the argument that drug quantity is an offense element under § 841(b), not a sentencing factor, was certainly available to Moss's counsel at the time of Moss's direct appeal. Our conclusion is consistent with other circuits which have spoken on the issue. See Sanders, 247 F.3d at 145-46; United States v. Smith, 241 F.3d 546, 548 (7th Cir. 2001); Garrott v. United States, 238 F.3d 903, 905-06 (7th Cir. 2001).

As far back as 1987, our circuit addressed the exact argument Moss now raises as a basis for relief, see United States v. Wood, 834 F.2d 1382 (8th Cir. 1987), and continued to address similar arguments throughout the early 1990s, see, e.g., United States v. Mabry, 3 F.3d 244, 250 (8th Cir. 1993), cert. denied, 511 U.S. 1020 (1994). In Wood, the defendant argued that his sentence under § 841(b)(1)(A) was invalid because drug quantity was not alleged in the indictment nor tried to the jury. See 834 F.2d at 1388. As our court said at that time, the argument presupposed that the enhanced penalty provisions available under § 841(b) were separate criminal offenses. Relying on McMillan v. Pennsylvania, 477 U.S. 79 (1986), we rejected Wood's argument after determining that Congress intended drug quantity to be a sentencing consideration, not an element of the offense. See Wood, 834 F.2d at 1390.

A barrage of similar arguments raged throughout the circuits in the late 1980s and early 1990s. See, e.g., United States v. Mergerson, 4 F.3d 337, 344 (5th Cir. 1993); United States v. Underwood, 982 F.2d 426, 429 (10th Cir. 1992); United States v. Lam Kwong-Wah, 966 F.2d 682, 685 (D.C. Cir. 1992); United States v. Perez, 960 F.2d 1569, 1574 (11th Cir. 1992); United States v. Trujillo, 959 F.2d 1377, 1381 (7th Cir. 1992); United States v. Lowden, 955 F.2d 128, 130 (1st Cir. 1992); United States v. Restrepo, 946 F.2d 654, 655 (9th Cir. 1991); United States v. Rigsby, 943 F.2d 631, 639-43 (6th Cir. 1991); United States v. Campuzano, 905 F.2d 677, 678-79 (2d Cir. 1990); United States v. Powell, 886 F.2d 81, 85 (4th Cir. 1989); United States v. Gibbs, 813 F.2d 596, 599 (3d Cir. 1987). The circuits, however, unanimously rejected the notion that drug quantity is an element of the offense. See United States v. Angle, 230 F.3d 113, 122 (4th Cir. 2000) (gathering cases). Although the argument was not rekindled by defense counsel until after Jones, the fact that it was raised extensively in the past, and explicitly addressed by this court previously, precludes a conclusion that the argument was "novel" and therefore unavailable because it was intellectually unascertainable.

Procedural default also cannot be overcome because the issue was settled in the lower courts. The Supreme Court has rejected the argument that default can be excused when existing lower court precedent would have rendered a claim unsuccessful. Bousley, 523 U.S. at 623 ("[F]utility cannot constitute cause if it means simply that a claim was unacceptable to that particular court at that particular time." (internal quotations omitted)).

In a somewhat analogous point, the dissent suggests, based on dictum in Reed v. Ross, 468 U.S. 1, 16 (1984), that cause may be shown where a new constitutional rule overturns "a longstanding and widespread practice to which [the Supreme Court] has not spoken, but which a near-unanimous body of lower court authority has expressly approved."11 The vitality of Reed has been questioned following the Supreme Court's decisions in Teague and Bousley. See, e.g., Simpson v. Matesanz, 175 F.3d 200, 212 (1st Cir. 1999); Boyer v. United States, 55 F.3d 296, 299 (7th Cir. 1995). Assuming arguendo that Reed remains valid, Apprendi does not fall within the exception relied upon by the dissent.

Reed suggests that a legal argument may be "unavailable" to counsel where contrary lower federal court authority has endured the test of time and there appears to be no discord among the courts on the issue. The origins of the exception are found in United States v. Johnson, 457 U.S. 537, 551 (1982), see Reed, 468 U.S. at 17 (citing Johnson), and two cases cited in Johnson as support for the proposition, see Gosa v. Mayden, 413 U.S. 665 (1973), and Stovall v. Denno, 388 U.S. 293 (1967). In both Gosa and Stovall, the Supreme Court refused to retroactively apply new rules affecting constitutional principles that had been followed by lower courts for over 100 years. See Gosa, 413 U.S. at 673, 685; Stovall, 388 U.S. at 299-300. The Supreme Court has never relied on the "longstanding and widespread practice" exception as a basis for excusing default, but based on its origin, the exception appears inapplicable when the issue has been settled for what is only a mere moment in the time line of lower federal court jurisprudence. The constitutional validity of judge-found quantity determinations was only conclusively established among the circuit courts in the early 1990s, and we respectfully disagree with the dissent's assertion that the lower courts' treatment of the issue constitutes a "longstanding" practice.

Finally, our conclusion that an Apprendi-type argument was reasonably available does not hold defense counsel to an unattainable or impractical standard of legal competence and sophistication. As we noted, defense counsel often challenged judge-found drug quantity determinations, and several commentators, and courts, had adequately set forth the legal basis supporting the proposition that drug quantity is an element of the offense. See, e.g., Susan N. Herman, The Tail That Wagged the Dog: Bifurcated Fact-Finding under the Federal Sentencing Guidelines and the Limits of Due Process, 66 S. Cal. L. Rev. 289 (1992); Judy Clarke, The Need for a Higher Burden of Proof for Factfinding under the Guidelines, 4 Fed. Sent. Rep. 300 (1992); Rigsby, 943 F.2d at 639-43 (following circuit precedent but explaining in detail why drug quantity under § 841(b) should be treated as an element of the offense). Because the Apprendi claim Moss now seeks to raise was reasonably available to his counsel, he cannot show cause for his failure to raise the issue as part of his direct appeal and is procedurally barred from raising it now.

For the foregoing reasons, we affirm the judgment of the district court.

RICHARD S. ARNOLD, Circuit Judge, dissenting.

The defendant in this case, Darius Moss, is now serving a sentence of 360 months (30 years) for conspiring to possess crack cocaine with the intent to distribute it, in violation of 21 U.S.C. §§ 841(a)(1) and 846. In fact, the statutory maximum for this offense, in Mr. Moss's circumstances, is 20 years. The United States has conceded, see ante at 996, that the sentence imposed violates the Constitution. This Court agrees. See ibid. Yet, the sentence is left in place, and Mr. Moss will serve ten years more than the Constitution allows. I cannot accept this state of affairs, and I therefore respectfully dissent.

In my view, the new rule of law announced in Apprendi v. New Jersey, 120 S. Ct. 2348 (2000), falls within the "watershed exception" to the non-retroactivity doctrine of Teague v. Lane, 489 U.S. 288 (1989). A rule that qualifies under this exception "must not only improve accuracy [of the trial and conviction], but also alter our understanding of the bedrock procedural elements essential to the fairness of a proceeding." Sawyer v. Smith, 497 U.S. 227, 242 (1990) (internal quotation marks and quoted cases omitted). Apprendi meets these qualifications. It raises the standard for determining factors that subject a criminal defendant to a higher term of imprisonment from a preponderance of the evidence to beyond a reasonable doubt, thereby increasing accuracy. It also requires such factors to be submitted to a jury, thereby enforcing a defendant's constitutional right to trial by jury. Similarly, the requirement that every element of a crime, defined as every fact that increases the statutory maximum, be charged in the indictment improves the accuracy of the fact-finding process, because it reduces the risk that an innocent person might be convicted of a more serious crime, or that a guilty person might be punished more severely than the law allows.

The language used by the Supreme Court itself in Apprendi is telling. The Court stated: "At stake in this case are constitutional protections of surpassing importance: the proscription of any deprivation of liberty without 'due process of law,' Amdt. 14, and the guarantee that '[i]n all criminal prosecutions, the accused shall enjoy the right to a speedy and public trial, by an impartial jury.' Amdt. 6." 120 S. Ct. at 2355. The Court described the state procedure before it, wherein a factor that increased the statutory maximum of an offense was decided by the judge, as "an unacceptable departure from the jury tradition that is an indispensable part of our criminal justice system," id. at 2366. The Court further recognized that the reasonable-doubt standard was at stake. This standard, as stated in In re Winship, 397 U.S. 358 (1970), "plays a vital role in the American scheme of criminal procedure. It is a prime instrument for reducing the risk of convictions resting on factual error. The standard provides concrete substance for the presumption of innocence--that bedrock axiomatic and elementary principle whose enforcement lies at the foundation of the administration of our criminal law." Id. at 363 (quoted case omitted).

To me, this signals the Supreme Court's own understanding that Apprendi recognizes bedrock procedures that are implicit in the concept of ordered liberty and that touch on the fundamental fairness of the trial. Indeed, four Justices almost explicitly endorsed a Teague exception for Apprendi claims by stating, "[t]oday, in what will surely be remembered as a watershed change in constitutional law, the Court imposes as a constitutional rule the principle it first identified in Jones." 120 S. Ct. at 2380 (O'Connor, J., dissenting). No doubt it is true that civilized systems of justice exist in which judges, not juries, decide criminal cases. But the Anglo-American tradition is otherwise. Every element of a criminal offense must be submitted to and found by a jury beyond a reasonable doubt. I find it regrettable that this Court feels free to minimize the hard-won right of trial by jury.

The Court makes the point that the "new rule announced [in Apprendi] floats and flows with the tide of legislative pronouncements." Ante, at 10. There is a sense in which this is true. It is certainly the case that legislatures, in this instance Congress, define the elements of crimes, and that these definitions are, at least to some degree, subject to change. This principle, however, is not without limits. Mullaney v. Wilbur, 421 U.S. 684 (1975), is a good example. In that case, the Supreme Court invalidated a Maine statute that presumed that a defendant who acted with an intent to kill possessed the "malice aforethought" necessary to constitute murder under state law. The statute purported to place on the defendant the burden of proving that he had acted, for example, in the heat of passion, so that he would be guilty of manslaughter instead of murder. The Supreme Court rejected the argument that the due-process and jury-trial protections expressed in Winship could be circumvented in this way. A state may not evade the right of trial by jury merely by "redefin[ing] the elements that constitute different crimes, characterizing them as factors that bear solely on the extent of punishment." Mullaney, 421 U.S. at 698. Thus, there are clearly some limits on Congress's ability to redefine elements of crimes so as to increase the punishment on the basis of judge-found facts. What these limits may be it is not necessary to explore in the present case. It suffices to say that the Apprendi rule does not subsist completely at the mercy of any and all congressional efforts to confine the right of jury trial only to some elements of crimes.

In addition, I question the Court's statement that an Apprendi violation is not a structural error requiring per se reversal. The Supreme Court, in Neder v. United States, 527 U.S. 1, 8-9 (1999), has held that it is not always reversible error to fail to submit to a jury an essential element of a criminal charge. If, for example, no reasonable jury could have found against the prosecution with respect to this element, the error can be treated as harmless. The same thing, however, cannot be said of the other part of the Apprendi principle -- that every element of a crime must be charged in an indictment (if the crime is federal). As far as I am aware, no case has ever held that the omission of an element of a crime from an indictment can be harmless error. In such cases, we do not ask whether a jury would have found that element on the evidence submitted to it, or, indeed, whether the grand jury would have returned an indictment including that element if it had been asked to do so. Rather, an indictment that omits an element of a crime is structurally deficient and provides no lawful basis for bringing anyone to trial. Failure to include an essential element in a federal indictment warrants relief even if the government later proves the omitted element at trial. See United States v. Zangger, 848 F.2d 923, 925 (8th Cir. 1988); United States v. Camp, 541 F.2d 737, 740 (8th Cir. 1976).

The Court also holds that Mr. Moss faces a procedural hurdle because he did not raise his Apprendi claim on direct appeal. The government relies on the rule stated in United States v. Bousley, 523 U.S. 614, 621-22 (1998) (quoted cases omitted), that collateral review

is an extraordinary remedy and 'will not be allowed to do service for an appeal.' . . . Where a defendant has procedurally defaulted a claim by failing to raise it on direct review, the claim may be raised in habeas only if the defendant can first demonstrate either 'cause' and actual 'prejudice,' or that he is 'actually innocent.'

I believe that this is an instance where a claim's legal basis was "not reasonably available to counsel," thereby establishing cause for failing to raise it on direct appeal. Reed v. Ross, 468 U.S. 1, 16 (1984), quoted in Bousley, 523 U.S. at 622. As explained by the Court in Reed v. Ross, such "cause" arises where a new constitutional rule overturns "a longstanding and widespread practice to which this Court has not spoken, but which a near-unanimous body of lower court authority has expressly approved." Id. at 17 (quoting United States v. Johnson, 457 U.S. 537, 551 (1982)). This is precisely the situation before us.

The rule announced in Apprendi was a departure from long accepted procedures. The dissenting opinion characterizes the holding as follows:

In its opinion, the Court marshals virtually no authority to support its extraordinary rule. Indeed, it is remarkable that the Court cannot identify a single instance, in the over 200 years since the ratification of the Bill of Rights, that our Court has applied, as a constitutional requirement, the rule it announces today.

120 S. Ct. at 2381. With this in mind, it cannot be fairly argued that an Apprendi claim was "reasonably available" to counsel at the time of Mr. Moss's appeal. Thus the failure to raise the Apprendi claim on direct appeal is excusable. Our Court today, citing other courts of appeals, questions the vitality of Reed, but the Supreme Court itself has reaffirmed Reed as recently as 1998. Bousley, supra, 523 U.S. at 622. In Reed, the novelty of a claim was held to excuse a lawyer's failure to raise it on direct appeal. It is ironic that the claim thus preserved from procedural default was a claim under Winship and Mullaney that a criminal defendant had been deprived of due process by an instruction that failed to require the prosecution to bear the burden of persuasion with respect to each element of a crime.