with whom The Chief Justice and Justice Scalia join, concurring.

I join the Court’s opinion and am in general agreement with its decision to apply the “plain statement” rule of Michigan v. Long, 463 U. S. 1032 (1983), to the state courts’ invocation of state procedural default rules. I write separately to emphasize two points. First, I do not read the Court’s opinion as addressing or altering the well-settled rule that the lower federal courts, and this Court, may properly inquire into the availability of state remedies in determining whether claims presented in a petition for federal habeas corpus have been properly exhausted in the state courts. See Humphrey v. Cady, 405 U. S. 504, 515-517 (1972); Ex parte Hawk, 321 U. S. 114, 118 (1944).

In 28 U. S. C. § 2254(b), Congress has provided that a writ of habeas corpus “shall not be granted unless it appears that the applicant has exhausted the remedies available in the courts of the State, or that there is either an absence of available State corrective processes or the existence of circumstances rendering such process ineffective to protect the rights of the prisoner.” The exhaustion requirement is not satisfied if the habeas petitioner “has the right under the law of the State to raise, by any available procedure, the question presented.” §2254(c). Thus, in determining wdiether a remedy for a particular constitutional claim is “available,” the federal courts are authorized, indeed required, to assess the likelihood that a state court will accord the habeas petitioner a hearing on the merits of his claim.

The rule requiring that a habeas petitioner exhaust available remedies in state court before seeking review of the same claims via federal habeas corpus serves- two important interests. First, its roots lie in the respect which the federal courts owe to the procedures erected by the States to correct constitutional errors, and the confidence that state court *269judges take, and should be encouraged to take, their constitutional duties seriously. Second, the rule furthers the interest in the efficiency of federal habeas corpus, by assuring that, in general, the factual and legal bases surrounding a petitioner’s constitutional claim or claims will have been developed in a prior adjudication. See generally Rose v. Lundy, 455 U. S. 509, 518-519 (1982).

To protect these interests we have held that where a federal habeas petitioner raises a claim which has never been presented in any state forum, a federal court may properly determine whether the claim has been procedurally defaulted under state law, such that a remedy in state court is “unavailable” within the meaning of § 2254(c). See Engle v. Isaac, 456 U. S. 107, 125-126, n. 28 (1982). The lower courts have consistently looked to state procedural default rules in making the “availability” determination, both before and after our decision in Engle. See, e. g., Watson v. Alabama, 841 F. 2d 1074, 1077, n. 6 (CA11), cert. denied, 488 U. S. 864 (1988); Leroy v. Marshall, 757 F. 2d 94, 97 (CA6), cert. denied, 474 U. S. 831 (1985); Wayne v. White, 735 F. 2d 324, 325 (CA8 1984); Williams v. Duckworth, 724 F. 2d 1439, 1442 (CA7), cert. denied, 469 U. S. 841 (1984); Richardson v. Turner, 716 F. 2d 1059, 1061-1062 (CA4 1983); Beaty v. Patton, 700 F. 2d 110, 112 (CA3 1983); Jackson v. Cupp, 693 F. 2d 867, 869 (CA9 1982); Matias v. Oshiro, 683 F. 2d 318, 319-321 (CA9 1982); Keener v. Ridenour, 594 F. 2d 581, 584 (CA6 1979); Smith v. Estelle, 562 F. 2d 1006, 1007-1008 (CA5 1977); United States ex rel. Williams v. Brantley, 502 F. 2d 1383, 1385-1386 (CA7 1974). Indeed, we have reaffirmed and applied the rule of Engle in Teague v. Lane, post, at 297-298.

A contrary rule would make no sense. It would require a “plain statement” indicating state reliance on a procedural bar where no state court was ever given the opportunity to pass on either the procedural posture or the merits of the constitutional claim. Moreover, dismissing such petitions *270for failure to exhaust state court remedies would often result in a game of judicial ping-pong between the state and federal courts, as the state prisoner returned to state court only to have the state procedural bar invoked against him. See Fay v. Noia, 372 U. S. 391, 435 (1963); United States ex rel Williams v. Brantley, supra, at 1385-1386 (“We refuse to contribute further needless and delaying requirements to a procedure that already often results in shuttling prisoners back and forth between the state and federal courts before any decision on the merits is ever reached”). Finally, such a rule would create an incentive to proceed immediately to federal court, bypassing state postconviction remedies entirely in the hope that the lack of a state court decision as to the applicability of the State’s procedural bar would be treated as “ambiguity.” Such a result would not only run counter to the decisions of this Court, see Rose, supra, at 518-519, but would also frustrate the congressional purpose embodied in § 2254.

In sum, it is simply impossible to “[r]equir[e] a state court to be explicit in its reliance on a procedural default,” ante, at 264, where a claim raised on federal habeas has never been presented to the state courts at all. In such a context, federal courts quite properly look to, and apply, state procedural default rules in making the congressionally mandated determination whether adequate remedies are available in state court.

My second concern stems from the majority’s references to our decisions in Murray v. Carrier, 477 U. S. 478 (1986), and Smith v. Murray, 477 U. S. 527 (1986). In these decisions, the Court reaffirmed the holding of Wainwright v. Sykes, 433 U. S. 72, 90-91 (1977), that a state prisoner pursuing federal habeas remedies must show both “cause” for a procedural default and “prejudice” flowing from the alleged constitutional violation for a federal court to entertain his claim on the merits despite the existence of an otherwise preclusive state-law ground for decision. In Murray v. Carrier, the Court rejected “a reworking of the cause and prejudice test ... to *271dispense with the requirement that the petitioner show cause and instead to focus exclusively on whether there has been a ‘manifest injustice’ or a denial of ‘fundamental fairness.’” 477 U. S., at 493. The Court went on to indicate that:

“We remain confident that, for the most part, ‘victims of a fundamental miscarriage of justice will meet the cause-and-prejudice standard.’ But we do not pretend that this will always be true. Accordingly, we think that in an extraordinary case, where a constitutional violation has probably resulted in the conviction of one who is actually innocent, a federal habeas court may grant the writ even in the absence of a showing of cause for the procedural default.” Id., at 495-496 (citation omitted).

At several points in its opinion, the Court refers to a “miscarriage of justice” test to be applied in conjunction with the cause and prejudice inquiry. See ante, at 258, and n. 2; ante, at 259, n. 3; ante, at 262. I do not read the Court’s opinion as suggesting any alteration of the relationship between the cause and prejudice inquiry and the narrow exception to the cause requirement where a petitioner cannot show cause but can make a strong showing of probable factual innocence. See Smith, supra, at 538-539 (“We similarly reject the suggestion that there is anything ‘fundamentally unfair’ about enforcing procedural default rules in cases devoid of any substantial claim that the alleged error undermined the accuracy of the guilt or sentencing determination”). The operative test is cause and prejudice; there is a kind of “safety valve” for the “extraordinary case” where a substantial claim of factual innocence is precluded by an inability to show cause. With this understanding, I join the Court’s opinion.

delivered the opinion of the Court.

In this case, we consider whether the ‘“plain statement’ rule” of Michigan v. Long, 463 U. S. 1032, 1042, and n. 7 (1983), applies in a case on federal habeas review as well as in a case on direct review in this Court. We hold that it does.

I

Petitioner Warren Lee Harris was convicted in the Circuit Court of Cook County, Ill., of murder. On direct appeal, petitioner challenged only the sufficiency of the evidence. The Appellate Court of Illinois, by an unpublished order, affirmed the conviction. App. 5; see 71 Ill. App. 3d 1113, 392 N. E. 2d 1386 (1979).

Petitioner then returned to the Circuit Court of Cook County and filed a petition for postconviction relief, alleging that his trial counsel had rendered ineffective assistance in several respects, including his failure to call alibi witnesses.1 The court dismissed the petition without an evidentiary hearing. The Appellate Court of Illinois, in another unpublished order, again affirmed. App. 9.

*258In its order, the Appellate Court referred to the “well-settled” principle of Illinois law that “those [issues] which could have been presented [on direct appeal], but were not, are considered waived.” Id., at 12. The court found that, “except for the alibi witnesses,” petitioner’s ineffective-assistance allegations “could have been raised in [his] direct appeal.” Ibid. The court, however, went on to consider and reject petitioner’s ineffective-assistance claim on its merits.

Petitioner did not seek review in the Supreme Court of Illinois. Instead, he pursued his ineffective-assistance-of-counsel claim in federal court by a petition for a writ of ha-beas corpus under 28 U. S. C. §2254. The District Court recognized that if the Illinois Appellate Court had held this claim to be waived under Illinois law, this Court’s decision in Wainwright v. Sykes, 433 U. S. 72 (1977), would bar a federal court’s consideration of the claim unless petitioner was able to show either “cause and prejudice” or a “miscarriage of justice.” 608 F. Supp. 1369, 1377 (ND Ill. 1985).2

The District Court, however, determined that the Illinois Appellate Court had not held any portion of the ineffective-assistance claim to have been waived. First, the District Court observed, the state court had “made clear” that the waiver did not apply to the issue of alibi witnesses. Id., at 1378. Second, the court never clearly held any other issue waived. The state court “did not appear to make two rulings in the alternative, but rather to note a procedural default and then ignore it, reaching the merits instead.” Ibid. Based on this determination, the District Court concluded that it was permitted to consider the ineffective-assistance claim in its entirety and ordered an evidentiary hearing. Id., at 1385. After that hearing, the court, in an unpublished *259memorandum and order, dismissed the claim on the merits, although it characterized the case as “a close and difficult” one. App. 45.

The Court of Appeals affirmed the dismissal, 822 F. 2d 684 (CA7 1987), but did not reach the merits because, in disagreement with the District Court, it believed the ineffective-assistance claim to be procedurally barred. Considering the Illinois Appellate Court’s order “ambiguous” because it contained “neither an explicit finding of waiver nor an expression of an intention to ignore waiver,” the Court of Appeals nonetheless asserted that a reviewing court “should try to assess the state court’s intention to the extent that this is possible.” Id,., at 687. Undertaking this effort, the Court of Appeals concluded that the order “suggested]” an intention “to find all grounds waived except that pertaining to the alibi witnesses.” Ibid. Based on this interpretation of the order, the Court of Appeals concluded that the merits of petitioner’s federal claim had been reached only “as an alternate holding,” ibid., and considered itself precluded from reviewing the merits of the claim.3

Concurring separately, Judge Cudahy stated: “Rather than attempting to divine the unspoken ‘intent’ of [the state] court, I think we should invoke a presumption that waiver not clearly found has been condoned.” Ibid.

The disagreement between the majority and the concurrence reflects a conflict among the Courts of Appeals over the standard for determining whether a state court’s ambiguous invocation of a procedural default bars federal habeas re*260view.4 We granted certiorari to resolve this conflict. 485 U. S. 934 (1988).

II

The confusion among the courts evidently stems from a failure to recognize that the procedural default rule of Wainwright v. Sykes has its historical and theoretical basis in the “adequate and independent state ground” doctrine. 433 U. S., at 78-79, 81-82, 87.5 Once the lineage of the rule is clarified, the cure for the confusion becomes apparent.

A

This Court long has held that it will not consider an issue of federal law on direct review from a judgment of a state court if that judgment rests on a state-law ground that is both “independent” of the merits of the federal claim and an “adequate” basis for the court’s decision. See, e. g., Fox Film Corp. v. Muller, 296 U. S. 207, 210 (1935); Murdock v. City of Memphis, 20 Wall. 590, 635-636 (1875). Although this doctrine originated in the context of state-court judgments *261for which the alternative state and federal grounds were both “substantive” in nature, the doctrine “has been applied routinely to state decisions forfeiting federal claims for violation of state procedural rules.” Meltzer, State Court Forfeitures of Federal Rights, 99 Harv. L. Rev. 1128, 1134 (1986).6

The question whether a state court’s reference to state law constitutes an adequate and independent state ground for its judgment may be rendered difficult by ambiguity in the state court’s opinion. In Michigan v. Long, 463 U. S. 1032 (1983), this Court laid down a rule to avoid the difficulties associated with such ambiguity. Under Long, if “it fairly appears that the state court rested its decision primarily on federal law,” this Court may reach the federal question on review unless the state court’s opinion contains a “ ‘plain statement’ that [its] decision rests upon adequate and independent state grounds.” Id., at 1042.7

The Long “plain statement” rule applies regardless of whether the disputed state-law ground is substantive (as it was in Long) or procedural, as in Caldwell v. Mississippi, 472 U. S. 320, 327 (1985). Thus, the mere fact that a federal claimant failed to abide by a state procedural rule does not, in and of itself, prevent this Court from reaching the federal claim: “[T]he state court must actually have relied on the procedural bar as an independent basis for its disposition of the *262ease.” Ibid. Furthermore, ambiguities in that regard must be resolved by application of the Long standard. Id., at 328.

B

The adequate and independent state ground doctrine, and the problem of ambiguity resolved by Long, is of concern not only in cases on direct review pursuant to 28 U. S. C. § 1257, but also in federal habeas corpus proceedings pursuant to 28 U. S. C. §2254.

Wainwright v. Sykes made clear that the adequate and independent state ground doctrine applies on federal habeas. 433 U. S., at 81, 87. See also Ulster County Court v. Allen, 442 U. S. 140, 148 (1979). Under Sykes and its progeny, an adequate and independent finding of procedural default will bar federal habeas review of the federal claim, unless the ha-beas petitioner can show “cause” for the default and “prejudice attributable thereto,” Murray v. Carrier, 477 U. S. 478, 485 (1986), or demonstrate that failure to consider the federal claim will result in a “‘fundamental miscarriage of justice.’” Id., at 495, quoting Engle v. Isaac, 456 U. S. 107, 135 (1982). See also Smith v. Murray, 477 U. S. 527, 537 (1986).

Conversely, a federal claimant’s procedural default precludes federal habeas review, like direct review, only if the last state court rendering a judgment in the case rests its judgment on the procedural default. See Caldwell v. Mississippi, 472 U. S., at 327; Ulster County Court v. Allen, 442 U. S., at 152-154. Moreover, the question whether the state court indeed has done so is sometimes as difficult to answer on habeas review as on direct review. Just as this Court under § 1257 encounters state-court opinions that are unclear on this point, so too do the federal courts under §2254.8

Habeas review thus presents the same problem of ambiguity that this Court resolved in Michigan v. Long. We held in *263Long that unless the state court clearly expressed its reliance on an adequate and independent state-law ground, this Court may address a federal issue considered by the state court. We applied that rule in Caldwell v. Mississippi, 472 U. S., at 327, to a “somewhat cryptic” reference to procedural default in a state-court opinion.

Although Long and Caldwell arose on direct review, the principles underlying those decisions are not limited to direct review. Indeed, our opinion in Caldwell relied heavily upon our earlier application of the adequate and independent state ground doctrine to habeas review in Ulster County. See Caldwell, 472 U. S., at 327-328. Caldwell thus indicates that the problem of ambiguous state-court references to state law, which led to the adoption of the Long “plain statement” rule, is common to both direct and habeas review. Faced with a common problem, we adopt a common solution: a procedural default does not bar consideration of a federal claim on either direct or habeas review unless the last state court rendering a judgment in the case “ ‘clearly and expressly’ ” states that its judgment rests on a state procedural bar. Caldwell, 472 U. S., at 327, quoting Long, 463 U. S., at 1041.9

C

Respondents, however, urge us to adopt a different rule for habeas cases, arguing that if a state-court decision is ambiguous as to whether the judgment rests on a procedural *264bar, the federal court should presume that it does. Respondents claim that applying the Long “plain statement” requirement to habeas cases would harm the interests of finality, federalism, and comity. This Court has been alert in recognizing that federal habeas review touches upon these significant state interests. Wainwright v. Sykes itself reveals this. See 433 U. S., at 90-91. We believe, however, that applying Long to habeas burdens those interests only minimally, if at all. The benefits, in contrast, are substantial.

A state court remains free under the Long rule to rely on a state procedural bar and thereby to foreclose federal ha-beas review to the extent permitted by Sykes.10 Requiring a state court to be explicit in its reliance on a procedural default does not interfere unduly with state judicial decision-making. As Long itself recognized, it would be more intrusive for a federal court to second-guess a state court’s determination of state law. 463 U. S., at 1041. Moreover, state courts have become familiar with the “plain statement” requirement under Long and Caldivell. Under our decision today, a state court need do nothing more to preclude habeas review than it must do to preclude direct review.

In contrast, respondents’ proposed rule would impose substantial burdens on the federal courts. At oral argument, counsel for respondents conceded that in some circumstances, under their proposal, the federal habeas court would be forced to examine the state-court record to determine *265whether procedural default was argued to the state court, or would be required to undertake an extensive analysis of state law to determine whether a procedural bar was potentially applicable to the particular case. See Tr. of Oral Arg. 28-29. Much time would be lost in reviewing legal and factual issues that the state court, familiar with state law and the record before it, is better suited to address expeditiously. The “plain statement” requirement achieves the important objective of permitting the federal court rapidly to identify whether federal issues are properly presented before it. Respondents’ proposed rule would not do that.11

Thus, we are not persuaded that we should depart from Long and Caldwell simply because this is a habeas case. Having extended the adequate and independent state ground doctrine to habeas cases, we now extend to habeas review the “plain statement” rule for determining whether a state court has relied on an adequate and independent state ground.12

*266hH HH

Applying the “plain statement” requirement in this case, we conclude that the Illinois Appellate Court did not “clearly and expressly” rely on waiver as a ground for rejecting any aspect of petitioner’s ineffective-assistance-of-counsel claim. Michigan v. Long, 463 U. S., at 1041. To be sure, the state court perhaps laid the foundation for such a holding by stating that most of petitioner’s allegations “could have been raised [on] direct appeal.” App. 12. Nonetheless, as the Court of Appeals recognized, this statement falls short of an explicit reliance on a state-law ground.13 Accordingly, this reference to state law would not have precluded our addressing petitioner’s claim had it arisen on direct review. As is now established, it also does not preclude habeas review by the District Court.

The judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings consistent with this opinion.

It is so ordered.

concurring.

In view of my dissent in Michigan v. Long, 463 U. S. 1032, 1065-1072 (1983), it is appropriate to add a few words explaining why there is unique virtue in applying the rule of that case to the problem presented by this case.

My dissent in Michigan v. Long was addressed primarily to two concerns. First, in adopting a presumption favoring the assertion of federal jurisdiction in ambiguous cases, the Court ignored the longstanding and venerated presumption *267that federal courts are without jurisdiction unless “ ‘the contrary appears affirmatively from the record.’” See Delaware v. Van Arsdall, 475 U. S. 673, 692 (1986) (Stevens, J., dissenting) (quoting King Bridge Co. v. Otoe County, 120 U. S. 225, 226 (1887)). Second, in its original form, the presumption adopted in Michigan v. Long expanded this Court’s review of cases in which state courts had overprotected their respective citizens. In my opinion, the federal courts — and •particularly this Court — have a primary obligation to protect the rights of the individual that are embodied in the Federal Constitution. See 475 U. S., at 695-697. Although some cases involving overly expansive interpretations of federally protected rights surely merit federal review, the interest in correcting such errors is necessarily secondary to the federal courts’ principal role as protector of federally secured rights. The expenditure of scarce judicial resources and the intrusion into state affairs is accordingly less justified when the state court has gone too far in protecting a federal right than when the state court has failed to provide the constitutional minimum of protection.

These concerns, however, are not implicated in a case such as this, in which a federal court, in considering a petition for a writ of habeas corpus, must decide whether a state procedural bar constitutes an adequate and independent state ground for denying relief. As our decisions in Fay v. Noia, 372 U. S. 391, 426-435 (1963), and Wainwright v. Sykes, 433 U. S. 72, 82-84 (1977), make clear, an adequate and independent state ground for decision does not dispossess the federal courts of jurisdiction on collateral review. More significantly, in considering petitions for relief under 28 U. S. C. § 2254, the federal courts do not review state-court decisions to determine if the States have gone too far in protecting the rights of their citizenry, but rather perform the core function of vindicating federally protected rights. Because the concerns that prevented me from joining the majority opinion in *268Michigan v. Long are not present in this case, I join the Court’s opinion and judgment.

dissenting.

This case presents the question whether a federal court may entertain a habeas corpus petition, without a showing of cause and prejudice, if the state court to which the federal *272claim was presented mentions a procedural default, yet considers also the merits of the claim. The majority holds that federal habeas courts must reach the merits of the federal issue absent explicit reliance on the bar, evidenced by a “plain statement” in the state court’s opinion.

Two premises underlie today’s holding. First, although the case before us is a federal habeas corpus proceeding, the majority explores whether an ambiguous reference to a state procedural bar would deprive us of jurisdiction in a matter here on direct review. The majority discovers that the rule of Michigan v. Long, 463 U. S. 1032 (1983), designed for cases where a state court explicitly relies on a state substantive ground that appears to be interwoven with federal law, applies as well in any direct review case where there is ambiguity concerning whether the state court intended to rely on a procedural bar. Thus fortified by its enhanced Long rule, the majority turns to the case before us. It stakes out the second premise, which is that direct review and collateral attack cases should be governed by the same rule. The majority therefore concludes that federal habeas courts must apply Long in determining whether a state court’s reference to a procedural bar triggers the cause-and-prejudice inquiry prescribed by Wainwright v. Sykes, 433 U. S. 72 (1977). Disagreement with each of the majority’s premises elicits my respectful dissent.

I

It is settled law that “where the judgment of a state court rests upon two grounds, one of which is federal and the other non-federal in character, our jurisdiction fails if the non-federal ground is independent of the federal ground and adequate to support the judgment.” Fox Film Corp. v. Muller, 296 U. S. 207, 210 (1935). The rule was first announced to deny our authority to revise state-court judgments resting on an alternative state substantive ground, e. g., Murdock v. Memphis, 20 Wall. 590, 636 (1875), and later was extended to bar our direct review of state judgments that rest on ade*273quate and independent state procedural grounds. See, e. g., Henry v. Mississippi, 379 U. S. 443, 446 (1965); Herndon v. Georgia, 295 U. S. 441 (1935). It follows that where a state court refuses to consider federal claims owing to a criminal defendant’s failure to comply with a state procedural rule that is otherwise adequate and independent, we lack authority to consider the claims on direct review.

Our discussions of this jurisdictional principle have identified circumstances where state procedural grounds are “inadequate” to support the result below, e. g., Johnson v. Mississippi, 486 U. S. 578, 587-589 (1988); James v. Kentucky, 466 U. S. 341, 348-349 (1984), and where state procedural grounds cannot be deemed “independent” of the underlying federal rights, e. g., Ake v. Oklahoma, 470 U. S. 68, 74-75 (1985). An analogous body of doctrine aids us in assessing the independence of state substantive grounds. See, e. g., Enterprise Irrigation Dist. v. Farmers Mutual Canal Co., 243 U. S. 157, 164 (1917). As might be expected in light of the common history and purposes of these doctrines, there is a significant degree of overlap. Our precedents declare, however, that “it is important to distinguish between state substantive grounds and state procedural grounds,” Henry v. Mississippi, 379 U. S., at 446, and caution against the indiscriminate application of principles developed in cases involving state substantive grounds to cases involving procedural defaults. See id., at 447. See also Wainwright v. Sykes, supra, at 81-82. These well-understood principles ensure our respect for the integrity of state-court judgments.

In Michigan v. Long, supra, we considered our jurisdiction to review a judgment of the Supreme Court of Michigan that had ruled a search unlawful. The state court’s opinion had relied almost exclusively on federal decisions construing the Fourth Amendment, though it twice cited an analogous state constitutional provision. 463 U. S., at 1043. After a review of our precedents considering whether various forms of references to state law constitute adequate and independent-state *274grounds, we adopted a presumption in favor of federal review “when ... a state court decision fairly appears to rest primarily on federal law, or to be interwoven with the federal law, and when the adequacy and independence of any possible state law ground is not clear from the face of the opinion . . . .” Id., at 1040-1041.

Our resolution of these ambiguities in favor of federal review rested on this critical assumption: When the state court’s judgment contains no plain statement to the effect that federal cases are being used solely as persuasive authority, and when state law is interwoven with federal law, we can “accept as the most reasonable explanation that the state court decided the case the way it did because it believed that federal law required it to do so.” Id., at 1041 (emphasis added). Our approach, we added, had the further advantage of not requiring us to interpret state laws with which we are generally unfamiliar. Id.., at 1039, 1041.

There may be a persuasive argument for applying Long to cases coming to this Court on direct review where the independence of a state procedural ground is in doubt because the state rule is interwoven with federal law. An example would be if “the State has made application of the procedural bar depend on an antecedent ruling of federal law, that is, on the determination of whether federal constitutional error has been committed.” Ake v. Oklahoma, supra, at 75. See also Longshoremen v. Davis, 476 U. S. 380, 388 (1986). But that situation is not presented in the case before us. In Illinois, “a defendant who neglects to raise a claim of inadequate representation on direct appeal may not later assert that claim in a petition for post-conviction relief,” United States ex rel. Devine v. DeRobertis, 754 F. 2d 764, 766, and n. 1 (CA7 1985) (collecting cases), though “strict application of [this] doctrine . . . may be relaxed . . . ‘where fundamental fairness so requires.’” People v. Gaines, 105 Ill. 2d 79, 91, 473 N. E. 2d 868, 875 (1984), cert. denied, 471 U. S. 1131 (1985), quoting People v. Burns, 75 Ill. 2d 282, 290, 388 N. E. 2d 394, 398 *275(1979). Petitioner does not claim that federal constitutional analysis is somehow determinative of “fundamental fairness” under Illinois law, or even that uncertainty exists on this point. Under the circumstances, “[t]here is no need for a plain statement indicating the independence of the state grounds since there was no federal law interwoven with this determination.” Pennsylvania v. Finley, 481 U. S. 551, 563 (1987) (Brennan, J., dissenting), citing Michigan v. Long, 463 U. S., at 1041.

The purported ambiguity in this case is much dissimilar from the ambiguity we confronted in Long.1 In Long “[t]he references to the State Constitution in no way indicated] that the decision below rested on grounds in any way independent from the state court’s interpretation of federal law,” id., at 1044 (emphasis in original), thus raising the question whether the state court “decided the case the way it did because it believed that federal law required it do so.” Id., at 1041. See also Pennsylvania v. Finley, supra, at 570 (Stevens, J., dissenting). The question in this case does not remotely implicate the independence of the state-law ground from federal law. The alleged ambiguity in the Illinois Appellate Court’s opinion relates instead to whether the state ground was invoked at all. The majority does not explain why adopting the Michigan v. Long presumption in this different context is sensible. It seems to me it is not.

*276Applied to this case, the “most reasonable explanation” test of Michigan v. Long suggests that the Illinois court referred to petitioner’s procedural default to rely on it, not because it was an interesting aside. The State’s rule is that failure to raise a claim on appeal is a waiver. The rule has an exception, presumably intended to apply to a smaller number of cases than the general rule of waiver, that operates to lift the procedural bar when justice so requires. Other States have adopted procedural default rules of like structure, see, e. g., Roman v. Abrams, 822 F. 2d 214, 222 (CA2 1987) (discussing analogous New York rule), cert, pending, No. 87-6154, and it may fairly be assumed that most procedural bars are in fact subject to some exception, even if a quite narrow one. There is no empirical or logical support, however, for the view that the most reasonable explanation for a court’s reference to the general rule is that the court intends to rely on some exception it does not mention. On the contrary, it is most unreasonable to adopt a rule that assumes either that state courts routinely invoke exceptions to their procedural bars without saying so, or that those courts are in the habit of disregarding their own rules.2

Indeed, if the majority’s aim is to devise a bright-line rule that will explain best the greatest number of similarly ambiguous state-court opinions, it should announce the mirror image of the rule adopted today. It should presume that the procedural bar was invoked unless the state court, by a “plain *277statement,” specifically relied on an exception. This alternative rule would serve the majority’s apparent concern for clarity in equal measure, and would be a far more accurate assessment of the intent of the state court in most cases. This rule would have the additional advantage of not presuming that a state court has disregarded its own laws in those instances where there is either no exception to the bar or an exception that manifestly is inapplicable to the defendant. Cf. Black v. Romano, 471 U. S. 606, 615 (1985) (“We must presume that the state judge followed [state] law”).

It is makeweight and unconvincing, moreover, to justify the majority’s extension of Michigan v. Long on the basis of our interest in avoiding unnecessary inquiries into “state laws with which we are generally unfamiliar.” Michigan v. Long, supra, at 1039.3 This concern is slight when the state-law ground is procedural rather than substantive. The doctrine of adequacy developed in the context of procedural bars already requires us to conduct extensive reviews of questions of state procedural law in order to determine whether the State’s “procedural rule is ‘strictly or regularly followed,’” Johnson v. Mississippi, 486 U. S., at 587, quoting Barr v. Columbia, 378 U. S. 146, 149 (1964), for state courts “may not avoid deciding federal issues by invoking procedural rules that they do not apply evenhandedly to all similar claims.” Hathorn v. Lovorn, 457 U. S. 255, 263 (1982). If this Court is institutionally capable of assessing whether a state procedural rule has been applied “evenhandedly to all similar claims,” it is certainly capable of assessing *278whether, in any given case, an exception to a procedural bar is applicable and has been invoked.4

The Court sidesteps the obvious difficulties of its new rule by stating that our decision in Caldwell v. Mississippi, 472 U. S. 320 (1985), already held that any ambiguity concerning whether a state court actually relied on a procedural bar “must be resolved by application of the Long standard.” Ante, at 262. It is true that Caldwell addressed the question whether the state court had relied on a procedural bar, and that it referred to Michigan v. Long in indicating, somewhat obliquely, that the lower court opinion did not contain an explicit statement that the decision was based on state law. 472 U. S., at 327. While Caldwell perhaps is not entirely clear on the point, it is difficult to view these statements as announcing conclusively that the Long presumption applies in all cases where there is doubt concerning whether a state court intended to rely on a procedural bar.

In any event, our references to the Long rule in Caldwell were entirely unnecessary to the decision, and the majority’s uncritical interpretation of Caldwell as controlling authority here is misplaced. In Caldivell two reasons persuaded us to reject the State’s argument that a procedural bar deprived us of jurisdiction. First, our own review of the state court’s opinion persuaded us that it could be “read . . . only as meaning that procedural waiver was not the basis of the decision.” Caldwell, supra, at 328 (emphasis added). Because we explicitly found that there was no ambiguity concerning whether the state court intended to rely on the procedural default, our references to Long ought not to be interpreted as requiring that Long be applied in cases where we are faced with such an ambiguity. Second, our opinion in Caldwell noted that Mississippi had not consistently applied *279its procedural bar to capital cases. 472 U. S., at 328. Caldwell’s citation of Michigan v. Long therefore cannot be characterized as holding that a procedural bar will oust this Court of jurisdiction only if the opinion below included the “plain statement” so eagerly sought by today’s majority. The State’s inconsistent application of its procedural bar would have rendered its bar inadequate in Caldwell, even if the state court had explicitly relied on it. See, e. g., Johnson v. Mississippi, supra, at 587-589. Caldwell ought not to be interpreted to require application of Long’s plain statement rule to the situation before us when the plainest possible statement could not have deprived us of jurisdiction in Caldwell itself.

I remain convinced that our reasoning in Michigan v. Long does not extend to a situation where, as here, there is doubt about whether a state court intended to rely on a procedural bar, but where there is no ambiguity, as there was in Long, concerning whether the bar is independent from federal law. Facial ambiguities that relate solely to whether a state court did invoke a procedural bar should not be resolved uncritically in favor of federal review.

II

Even if the majority were correct in concluding that the judgment of the Illinois Appellate Court would have been reviewable in this Court under Michigan v. Long, it errs in concluding that federal habeas review must also be available. The equivalence the majority finds between direct and collateral review appears to be based on two arguments. First, the majority asserts that Wainwright v. Sykes, 433 U. S. 72 (1977), “made clear that the adequate and independent state ground doctrine applies on federal habeas.” Ante, at 262. Second, the Court argues that the “substantial” benefits of extending Michigan v. Long to the habeas context outweigh any state interests that may be burdened by applying Long in this context. Neither argument is persuasive.

*280Far from supporting the majority’s reflexive extension of Long to habeas cases, Wainwright v. Sykes made clear, after an exhaustive review of our precedents, that the adequate and independent state ground doctrine does not “apply” in the habeas context in the manner suggested by the Court today. As Sykes noted, our decision in Fay v. Noia, 372 U. S. 391 (1963), explicitly divorced the doctrines governing our appellate jurisdiction from those governing the power of the federal courts to entertain habeas corpus applications. Wainwright v. Sykes, supra, at 82-83; Fay v. Noia, 372 U. S., at 425-426, 433-434. Under the view we took of the habeas corpus statute in Fay, the state court’s reliance on its procedural rule, even if sufficient to preclude direct review of the state-court judgment, could not prevent a federal habeas court from considering the underlying constitutional claim. It was only as a matter of comity that we recognized the principle that habeas review could be denied to an “applicant who ha[d] deliberately by-passed the orderly procedure of the state courts and in so doing ha[d] forfeited his state court remedies.” Id., at 438.

Our decision in Sykes placed some limits on the expansive regime of Fay v. Noia, but reaffirmed that comity and federalism are the principles that control the weight that a federal habeas court should accord to a state procedural default. These constitutional concerns, not some mechanical application of the doctrines governing our appellate jurisdiction, formed the basis for our holding that a state procedural default will preclude federal habeas review unless the applicant shows both cause for failing to comply with the State’s rule and actual prejudice resulting from the alleged constitutional violation.5 433 U. S., at 84-91. Indeed, the majority’s re*281affirmation of the authority of federal courts to grant habeas relief, notwithstanding a procedural default, on a showing of “cause and prejudice” belies any facile equivalence between direct and collateral review. The significance of Sykes for this case has nothing to do with “adequate and independent state grounds,” but with principles governing the relationship between federal and state courts that have become an essential part of our judicial federalism.

Because our decision to honor state procedural defaults in habeas cases is intended “to accord appropriate respect to the sovereignty of the States in our federal system,” Ulster County Court v. Allen, 442 U. S. 140, 154 (1979), any determination that a state court did not intend to rely on a procedural default must be made with the same deference to the State’s sovereignty that motivates our willingness to honor its procedural rules in the first place. The majority’s second argument for extending Michigan v. Long to the habeas context seems to acknowledge as much, for at least it purports to be guided by those principles of federalism and comity that until now have informed our analysis. Ante, at 263-265. The majority’s perfunctory discussion of these principles, however, is inadequate to justify its view that extending Long will burden state interests “minimally, if at all,” ante, at 264, while producing “substantial” benefits. Ibid. These conclusions, in my view, reflect a miserly assessment of the State’s interest and an extravagant notion of the benefits to be derived from extending Long to habeas cases.

The majority dismisses the State’s interests by positing that state courts have become familiar with the “plain statement” rule under Long. One may question whether it is not *282“unrealistic — and quite unfair — to expect the judges [who must deal with postconviction proceedings in the lower state courts] to acquire and retain familiarity with this Court’s jurisprudence concerning the intricacies of our own jurisdiction.” Pennsylvania v. Finley, 481 U. S., at 570 (Stevens, J., dissenting). In any event, the majority’s improvident extension of Michigan v. Long burdens significant state interests that today’s opinion does not even acknowledge. As we emphasized at great length in Engle v. Isaac, 456 U. S. 107, 126-128 (1982), federal habeas review itself entails significant costs. It disturbs the State’s significant interest in repose for concluded litigation, denies society the right to punish some admitted offenders, and intrudes on state sovereignty to a degree matched by few exercises of federal judicial authority. The majority’s new rule can only increase the likelihood that these costs will be incurred more often.

The majority’s opinion also reflects little consideration of the perverse incentives created by its holding. Because an ambiguous state-court order will ensure access to a federal habeas forum, prisoners whose claims otherwise would be procedurally barred now have every incentive to burden state courts with a never-ending stream of petitions for post-conviction relief. Such perseverance may, in due course, be rewarded with a suitably ambiguous rebuff, perhaps a one-line order finding that a prisoner’s claim “lacks merit” or stating that relief is “denied.” Instead of requiring prisoners to justify their noncompliance with state procedural rules, as contemplated by the cause-and-prejudice standard, the majority’s decision openly encourages blatant abuse of state-court processes and circumvention of the standard established in Sykes.6

*283The majority’s explanation of the questionable advantages of its new rule does not allay these concerns. The majority appears to think that state procedural rules are so arcane that the federal district courts and courts of appeals should not be burdened with the task of determining their controlling effect. We have recognized, however, that those courts are experts in matters of local law and procedure. See, e. g., Rummel v. Estelle, 445 U. S. 263, 267, n. 7 (1980) (deferring to the Fifth Circuit’s conclusion that petitioner was not procedurally barred under Texas law); Ulster County Court v. Allen, supra, at 153-154 (noting deference owed to the Second Circuit’s conclusion that New York court decided constitutional issue on the merits); Brown v. Allen, 344 U. S. 443, 458 (1953) (“So far as weight to be given the proceedings in the courts of the state is concerned, a United States district court, with its familiarity with state practice is in a favorable position to recognize adequate state grounds in denials of relief by state courts without opinion”). Indeed, far from regarding decisions of state-law questions as “substantial burdens on the federal courts,” ante, at 264, our precedents reveal that a federal court’s ability to dispose of cases on state-law grounds is an affirmatively desirable means of avoiding, if possible, federal constitutional questions. See, e. g., Pennhurst State School and Hospital v. Halderman, 465 U. S. 89, 117, 119, n. 28 (1984); Ashwander v. TVA, 297 U. S. 288, 347 (1936) (Brandeis, J., concurring); Delaware v. Van Arsdall, 475 U. S. 673, 693, n. 5 (1986) (Stevens, J., dissenting). Our limited familiarity with local law may require some relaxation of this salutary principle in this Court, *284but the majority offers no sound reason for thinking that the other federal courts are in dire need of such a dispensation, especially when it is conferred at the cost of an undetermined increase in the number of cases to be resolved on the merits.

Even assuming that avoidance of state-law questions is now considered an unalloyed blessing as a general matter, those questions cannot be avoided in federal habeas cases. To cite only the most obvious reason, the habeas statute and our decisions preclude habeas relief “unless it appears that the applicant has exhausted the remedies available in the courts of the State, or that there is either an absence of State corrective process or the existence of circumstances rendering such process ineffective to protect the rights of the prisoner.” 28 U. S. C. § 2254(b). See Granberry v. Greer, 481 U. S. 129, 133-134 (1987); Picard v. Connor, 404 U. S. 270, 275 (1971); Ex parte Hawk, 321 U. S. 114, 116-118 (1944). Our cases recognize that this requirement refers only to remedies still available at the time of the federal habeas petition, and that no such remedies are in fact available if the state courts would refuse to entertain the claim because of a procedural default. Engle v. Isaac, supra, at 125-126, n. 28. Thus, federal habeas courts must become familiar with state rules governing procedural defaults and with the. circumstances when exceptions to these rules will be invoked. Because the unequivocal command of § 2254(b) already requires that federal courts become experts on the procedural rules that govern the availability of postconviction relief in the state courts, the majority’s assessment of the marginal burdens imposed on federal courts by the need to construe those rules in cases like the one before us can only be described as extravagant.

Our decision in Engle v. Isaac, supra, which the Court strongly reaffirms in this case and in two other cases decided today, ante, at 263, n. 9; Castille v. Peoples, post, p. 346; Teague v. Lane, post, p. 288, thus belies the majority’s assessment of the benefits of its new rule. Engle also *285indicates that there is a difficulty even more fundamental with the majority’s reasoning. The majority’s premise, indeed the driving force of its holding, appears to be that there is always a possibility that the state courts will forgive a procedural default irrespective of how clear state law may be to the contrary. But this premise is not credited even by the majority. If this premise were true, Engle would be overruled, not reaffirmed. If forgiveness were always a realistic possibility, no federal habeas court could ever invoke Engle, for no federal court could be sure, in any given case, that the state courts would refuse to consider a federal claim on the basis of the state’s procedural default rules.

According to the majority, two different rules will guide the lower courts’ consideration of procedural default issues after today. On the one hand, if a defendant presents his claims to the courts of the State, the majority’s new rule applies. A federal habeas court faced with an ambiguous state-court opinion may not consult state-law sources to determine whether the state court is authorized to forgive the procedural default, or to decide whether the circumstances in which a default may be overlooked consistent with state law are present in the particular case. On the other hand, if a defendant has never attempted to raise his claim in the courts of the State, Engle applies. A federal habeas court faced with such a case must look to state law to decide whether the petitioner is procedurally barred and whether the state courts are likely to waive his procedural default. The federal court must apply our holding in Wainwright v. Sykes, 433 U. S. 72 (1977), if the court concludes, on the basis of such review, that the state courts would likely refuse to entertain the petitioner’s claim. Yet it is obvious that Engle and the rule adopted by the majority in this case are based on irreconcilable assumptions about the regularity and predictability of state procedural rules. And it is not difficult to predict that the lower courts, faced with inconsistent pro*286nouncements from this Court, soon will require us to choose one principle or the other.

Nothing could illustrate this point better than the Court’s decision in Teague v. Lane, post, p. 288. The petitioner in Teague, like Harris, failed to raise one of his federal constitutional claims on direct appeal in the Illinois courts. Under the same Illinois procedural rule at issue in the present case, the petitioner in Teague is barred from raising his claim in collateral proceedings unless fundamental fairness requires that his default be overlooked. Speaking for four Members of the Court, Justice O’Connor concludes that the petitioner in Teague has exhausted his state remedies because, in view of the limited scope of the fundamental fairness exception, the Illinois courts clearly would refuse to entertain the merits of his claim in collateral proceedings. For the same reason, the Teague plurality concludes that the petitioner in that case is procedurally barred. Teague v. Lane, post, at 297-299. Without disagreeing with the plurality’s conclusion on these logically antecedent issues, Justice White concurs in the judgment disposing of the case on retroactivity grounds. Post, at 317. It appears therefore that five Members of the Court are of the view that it would be entirely futile to remand the case to the Illinois courts because those courts enforce their procedural default rules strictly. The majority does not explain, and I fail to see, how this conclusion can possibly be squared with the majority’s adoption of a conclusive presumption to the contrary in the present case.

In sum, the Court’s decision to extend Michigan v. Long to the habeas context ignores important state interests that it is our tradition to honor, and advances no significant federal interest. Indeed, the Court’s new rule works against the important federal interests of avoiding, if possible, decisions on federal constitutional claims, and stemming the overwhelming tide of prisoner petitions. Neither logic nor precedent requires this perverse result.

*287h — 1 H-t

The Court of Appeals acknowledged that petitioner properly preserved for federal review the claim that his counsel was ineffective in failing to call alibi witnesses. However, the Court of Appeals failed to address the merits of this claim. Nor did the court inquire whether, with respect to those claims that the court determined to be procedurally barred, petitioner could establish cause and prejudice and thus secure federal habeas review. I would vacate the judgment of the Court of Appeals and remand for further consideration of these matters. Because the Court’s remand goes significantly further, I dissent.