I. INTRODUCTION

Claims that errors were made during jury selection are among the most common of all grounds for criminal appeals. (1) Yet appellate courts, both state and federal, seem profoundly confused about how to analyze these kinds of errors. The confusion has its roots in a defendant's

ability to cure most jury selection errors with the use of a peremptory challenge, and the resulting question of whether the price of such a cure--the loss of a peremptory challenge--is itself an injury requiring reversal.

Consider, for example, two state court cases that reached opposite conclusions about errors that led to the loss of a defendant's peremptory challenge. The first case is from Vermont, where John Doleszny was charged with sexual assault on a victim under sixteen. (2) During jury selection, a prospective juror disclosed that he was acquainted with one of the prosecution witnesses, a doctor who had examined the alleged victim after the assault. When asked if he could be impartial in evaluating the testimony of the doctor, the prospective juror replied, "I certainly could try to be impartial but I'm not saying that I could." (3) The prosecution was unable to rehabilitate this response. (4)

Despite the prospective juror's inability to assure the trial judge that he could be impartial, the judge erroneously denied Mr. Doleszny's challenge for cause. Mr. Doleszny was then forced to use a peremptory challenge to remove the prospective juror. Since Mr. Doleszny eventually exhausted his peremptory challenges, he was in effect deprived of one peremptory challenge as a result of the trial judge's erroneous denial of his challenge for cause.

On appeal, the Vermont Supreme Court reversed the conviction and remanded for a new trial. Such an error, said the court, is automatic reversible error. (5) There are many jurisdictions that analyze these kinds of errors the same way: if the defense lost a peremptory challenge as a result of the judge's error, the conviction is automatically reversed without further inquiry. (6)

But other jurisdictions would apply a harmless error analysis to the same error. (7) One example is a case from Idaho, where Eric Ramos was charged with two counts of lewd conduct with minors. (8) During jury selection, a prospective juror expressed his belief that the defendant "must have done something" to be on trial. (9) As in Doleszny, the prosecution was unable to rehabilitate this prospective juror. (10) When the trial judge refused to remove him for cause, Mr. Ramos used a peremptory challenge to remove the putative juror and eventually exhausted his peremptory challenges. (11)

Instead of stopping at the cause error itself and simply reversing, the Idaho Supreme Court examined the impact of the error on Mr. Ramos' trial, specifically, whether there was anything wrong with the jury that had convicted him. Because Mr. Ramos removed the biased juror with one of his peremptory challenges and there was no evidence that the jury that convicted him had not been fair and impartial, the Idaho Supreme Court concluded that Mr. Ramos had not been prejudiced by the trial judge's erroneous cause ruling and affirmed the conviction. (12)

These two cases represent two very different approaches to a very common problem. While the cases often involve erroneous denials of a defense challenge for cause, as in Doleszny and Ramos, there are many other situations in which a variant of the same problem commonly arises. For example, when a trial judge erroneously grants a prosecution challenge for cause and the prosecution exhausts all of its peremptory challenges, that error has the effect of giving the prosecution an additional peremptory challenge. Like Doleszny and Ramos, defendants in such cases suffer an imbalance in peremptory challenges, although the source of the imbalance is that the prosecution received more peremptory challenges rather than that the defendant received fewer. Is such an error automatically reversible, or could it be harmless?

In this article, we begin by reviewing the Supreme Court's three opinions in the area of "curative peremptory challenges," including its most recent pronouncement in United States v. Martinez-Salazar. (13) We then discuss the historical relationship between challenges for cause and peremptory challenges, arguing that peremptory challenges have always had a curative purpose. Drawing on that history and on the evolution of the doctrine of harmless error, we break down the types and permutations of these kinds of jury selection errors into six different scenarios. (14) We argue that in each of the six scenarios the resulting imbalance in peremptory challenges should not result in reversals, both because the error is harmless and, at least in some of the scenarios, because the defendant should be deemed to have waived the error. (15)

II. AN UNCERTAIN SUPREME COURT: THE ABOUT FACE FROM GRAY V. MISSISSIPPI TO ROSS V. OKLAHOMA

A. Gray v. Mississippi

The Supreme Court has not tackled the problem of curative peremptory challenges very often, and on those few occasions when it has addressed the issue it has not given very clear guidance. Its initial foray was in a 1987 state death penalty case, Gray v. Mississippi. (16) In that case, the trial judge erroneously removed a prospective juror for cause on the prosecution's motion. (17) The Mississippi Supreme Court affirmed the conviction and death sentence, despite numerous irregularities in the jury selection, (18) concluding that any cause error that may have resulted in an unwarranted increase in the number of prosecution peremptory challenges was harmless. (19)

The United States Supreme Court reversed, rejecting the prosecution's harmless error argument. Although the prosecution insisted that it would have peremptorily struck the last prospective juror if it had had any peremptory challenges left, the Court noted that the timing of a trial judge's error in a particular case could very well affect the prosecution's complex strategic decisions about when and on whom to expend its remaining peremptory challenges. (20) The Court concluded that the "unexercised peremptory argument" that underlay the prosecution's harmless error contention mischaracterized the nature of the error. It framed the question as being not whether a cause error resulted in the erroneous exclusion of a particular prospective juror, but rather whether "the composition of the jury panel as a whole could possibly have been affected by the trial court's error." (21) It concluded that it could not say that the cause errors suffered by Mr. Gray could not have "possibly" affected the composition of the jury as a whole, and it therefore held that the error was automatically reversible. (22) Interestingly, the Court reversed only Mr. Gray's death sentence, not his conviction.

Gray seemed like powerful support for the notion that virtually all jury selection errors should be automatically reversed. If the Court will reverse any conviction where the composition of the jury panel "could possibly have been affected by the trial court's error," (23) then there is a far stronger argument for automatic reversal in cases like Doleszny and Ramos, where there is no doubt that the composition of the actual juries would have been different had the defense not lost a peremptory challenge due to the trial judge's cause error.

Gray was a strong opinion, but the Court was badly split. Four dissenters argued that the problem jurors should have been excluded in any event, and thus that the error would not have affected the composition of the jury even under the majority's expansive recognition of systemic jury selection errors and its low "possibility" threshold. (24) Perhaps because the Court was so deeply divided, it undertook the rather remarkable step of reexamining the issue less than a year later, in Ross v. Oklahoma. (25)

B. Ross v. Oklahoma

Ross was another death penalty case. Unlike in Gray, however, where the imbalance in peremptory challenges was the result of the trial judge erroneously excluding qualified jurors, the peremptory imbalance in Ross arose from the trial judge erroneously failing to exclude a disqualified juror. Specifically, the trial judge in Ross failed to apply Witherspoon-Witt to excuse a prospective juror who expressed disqualifying pro-death penalty views. The prospective juror in question indicated during voir dire that if the jury found the defendant guilty he would automatically vote to impose the death penalty regardless of any mitigating factors. (26) This sort of reverse Witherspoon-Witt attitude entitled the defense to remove the prospective juror for cause. (27) When the trial court erroneously denied Mr. Ross's challenge for cause, he was forced to use one of his peremptory challenges to remove the pro-death penalty juror. He eventually exhausted his peremptory challenges. Thus, as in Doleszny and Ramos, the effect of this error was that the defense ended up with one fewer peremptory challenge than guaranteed by state law. (28)

The dissenters in Gray, now a majority with the loss of Justice Powell and the addition of Justice Kennedy, declined to apply Gray's automatic reversal rule. Instead, the Court announced that "the broad language used by the Gray Court is too sweeping to be applied literally, and is best understood in the context of the facts there involved." (29) Gray, said the new Ross majority, must be limited to its context and applied only to the erroneous exclusion of a qualified juror in a capital case, not to the erroneous inclusion of a disqualified juror in a capital case who is later peremptorily excused. (30)

Freed from the automatic reversal rule of Gray, the Ross Court proceeded to fashion a rather startling syllogism to conclude that Mr. Ross suffered no error at all, let alone reversible error. It began by noting that under Oklahoma law, as in virtually all other states, (31) a defendant waives the right to argue that a trial judge erroneously denied a defense challenge for cause if the defendant fails to use an available peremptory challenge to remove the erroneously retained juror. (32) From this waiver rule, the Court reasoned that Oklahoma intended that the nine peremptory challenges it chose to give capital criminal defendants necessarily include any peremptory challenges that a defendant might have to use curatively. (33) Thus, it concluded that Mr. Ross "received all [the peremptory challenges] that Oklahoma law allowed him," (34) and therefore that there was no error at all in the manner Mr. Ross's jury was selected, let alone reversible error.

The four-justice remnant of the Gray majority filed a vociferous dissent in Ross, accusing the majority of "unaccountably" departing from the one-year precedent of Gray. (35) It argued that Mr. Ross suffered an even more direct injury to his "right" to peremptory challenges than did Mr. Gray. The imbalance in peremptory challenges in Ross was caused by Mr. Ross being forced to expend one of his peremptory challenges to cure the trial judge's erroneous denial of his challenge for cause; the imbalance in Gray was caused by the arguably much less significant decision by the trial court to grant, erroneously, several prosecution challenges for cause. If Mr. Gray was deprived of due process because the prosecution in his case in effect received a few more peremptory challenges than allowed by law, surely Mr. Gray was deprived of exactly the same due process when he was forced to use a peremptory challenge on a prospective juror who should have been excused for cause.

Gray and Ross were decided according to state jury selection procedures. In a footnote, the Ross majority specifically left open the question of whether Federal Rule of Criminal Procedure 24(b), like its Oklahoma counterpart, requires peremptory challenges to be used curatively and therefore whether the Ross "no error" analysis will apply to federal cases. (36) Of course, Gray and Ross were also both death penalty cases, raising the issue of whether they should be limited to the unique jury selection challenges presented in capital cases. (37) The Court addressed these unresolved issues twelve years later, in United States v. Martinez-Salazar. (38)

III. UNITED STATES V. MARTINEZ-SALAZAR

A. The Jury Selection Error

Abel Martinez-Salazar and a co-defendant were indicted in the United States District Court for the District of Arizona on charges of conspiracy to sell heroin, drug trafficking in heroin, and carrying a firearm in the course of the drug trafficking. (39) As in Ross and as in our examples, Doleszny and Ramos, the jury selection error that took Martinez-Salazar to the Supreme Court began with the failure of the trial judge to remove a prospective juror for cause. In response to a question on a jury questionnaire asking if he knew of anything that might affect his ability to be impartial, the prospective juror wrote, "I would favor the prosecution." (40) When the venire was brought into court, the trial judge asked the prospective juror several questions about his response on the questionnaire. Most of the replies were somewhat ambiguous. (41) But at one point in the course of his colloquy with the judge the prospective juror said, "I think what I am saying is all things being equal, I would probably tend to favor the prosecution." (42)

Mr. Martinez-Salazar and his co-defendant challenged the prospective juror for cause, but the trial judge erroneously denied the challenge. (43) The defendants then used one of their eleven shared peremptory challenges to strike the juror, and later exhausted all of their peremptory challenges. (44) Thus, as in Ross, the impact of the trial judge's error was that the defense had one fewer peremptory challenge than it would otherwise have had. (45) Both defendants were convicted, and Mr. Martinez-Salazar appealed, arguing both a general insufficiency of evidence and the trial court's erroneous refusal to grant the challenge for cause. (46)

The Ninth Circuit reversed the conviction. It agreed with the prosecution that under Ross there was no Sixth Amendment violation, since the jury that convicted Mr. Martinez-Salazar was fair and impartial. (47) Nevertheless it reversed the conviction, concluding that the impairment of Mr. Martinez-Salazar's right to a full complement of peremptory challenges denied him due process. (48)

B. The Supreme Court's Strained Analysis

The federal circuits were split on the issue of whether a curative loss of a peremptory challenge was reversible. (49) By granting certiorari in Martinez-Salazar, the Court had an opportunity not only to resolve this split among the circuits, but also to address the two more general questions left open by Ross--whether Federal Rule 24(b) requires peremptory challenges to be used curatively and whether the principles in Gray/Ross have any application at all in non-capital cases. It was also an opportunity to provide much-needed guidance for state courts, which were likewise deeply divided on the question of how to analyze these kinds of jury selection errors. (50)

The government argued that the Court should read Federal Rule 24(b) as requiring the curative use of peremptory challenges, and that under such a reading the Ross syllogism would be dispositive. But the Court refused to take that approach. In an opinion written by Justice Ginsburg, in which five other Justices joined and the three others specially concurred, the Court held that Federal Rule 24(b), unlike its Oklahoma counterpart, does not mandate that peremptory challenges be used curatively. (51) Thus, it concluded that Mr. Martinez-Salazar, unlike Mr. Ross, was in fact entitled to the full federal compliment of eleven peremptory challenges.

But the Court then turned that entitlement on its head, by reasoning that precisely because Mr. Martinez-Salazar was not forced to use any peremptory challenges curatively, he could not complain about his voluntary election to do so. He was entitled to, and received, all of his peremptory challenges, so there was no error. (52) Although the Court acknowledged that the trial judge had put the defendants in a difficult position by erroneously denying their challenge for cause, the Court reasoned that the defendants had a choice at that point of whether or not to remove the biased juror peremptorily. The defendants, according to the Court, "had the option of letting [the juror] sit on the petit jury and, upon conviction, pursuing a Sixth Amendment challenge on appeal." (53) Unattractive as this would have been, the Court opined, "[a] hard choice is not the same as no choice." (54)

Justice Scalia, in a concurring opinion joined by Justice Kennedy, objected to the majority's assertion that Mr. Martinez-Salazar's lawyer could have allowed the biased juror to remain and then appealed the conviction. Besides pointing out that this issue was not before the Court, Justice Scalia argued that "normal principles of waiver ... disable a defendant from objecting on appeal to the seating of a juror he was entirely able to prevent." (55) Justice Scalia also suggested that "it may well be regarded as one of the very purposes of peremptory challenges to enable the defendant to correct judicial error on [juror-bias questions]." (56) In the next section on the history of peremptory challenges and their relationship to challenges for cause, we will show that Justice Scalia's suggestion is indeed correct. (57)

C. The Court's Fear of Harmless Error

What is striking about the Ginsburg and Scalia opinions in Martinez-Salazar is the extent to which they strain, in their own ways, to avoid reaching the harmless error issue that we contend is at the heart of the problem of curative peremptory challenges. In doing so, both stretch themselves into some very difficult contortions. (58)

Justice Scalia concurs in the majority's flawed "no error" syllogism, yet chides the majority for reaching the waiver issue. But the majority's "no error" conclusion is entirely dependent on its "no waiver" conclusion. It is only after the majority interprets Federal Rule 24(b) as not requiring the curative use of peremptory challenges that it can engage in the fiction that Mr. Martinez-Salazar was not "forced" to expend a peremptory challenge to cure the cause error, and that "a hard choice is not the same as no choice."

The majority opinion is more problematical. By holding that a criminal defendant is not required to cure cause errors with peremptory challenges, the majority turns well-settled principles of waiver upside down. A fundamental tenet of judicial review is the "contemporaneous objection role," under which a lawyer must object to all trial errors during trial, other than those few denominated as "plain" errors, in order to preserve the errors on appeal. (59) What is different about peremptory challenges, of course, is that the defense has at its disposal an independent mechanism to correct an erroneous failure to remove a juror for cause. Yet the Martinez-Salazar Court suggests that a defendant can elect not to cure the cause error and nevertheless raise it on appeal. As Justice Scalia points out, such a suggestion is inimical to the contemporaneous objection rule and the axiom of waiver upon which it is built:

   The difficult question ... is not whether Federal Rule of Criminal
   Procedure 24(b) requires exercise of the peremptory, but rather whether
   normal principles of waiver (not to say the even more fundamental principle
   of volenti non fit injuria) disable a defendant from objecting on appeal to
   the seating of a juror he was entirely able to prevent. (60)

The Martinez-Salazar Court's refusal to read basic and well-settled notions of waiver into Rule 24(b) was particularly mysterious because it led the Court to engage in a troubling line of reasoning about the role of defense counsel. We cannot imagine a circumstance in which a competent defense lawyer could ethically, or would practically, decide to infect his client's jury with a demonstrably biased juror who could have been removed with an available peremptory challenge. Imagine the logic of such a decision: "I think I'll keep that Klansman on the jury, who said in open court that he could not be fair to my black defendant, because I want to keep my peremptory challenges available so I can strike all Tauruses, who are well known to be pro-prosecution." No competent defense lawyer would ever trade away the opportunity to excuse a demonstrably biased juror for the chance to peremptorily excuse jurors he or she has an inarticulable hunch might be biased.

Moreover, as a practical matter, how often will a defense lawyer be so sure of the righteousness of his challenge for cause that he will gamble everything on a reversal based on that challenge? As the majority itself points out in Martinez-Salazar: "Challenges for cause and rulings upon them ... are fast paced, made on the spot and under pressure. Counsel as well as court, in that setting, must be prepared to decide, often between shades of gray, 'by the minute."' (61)

Precisely because black shades into gray, and because "obvious" challenges for cause are seldom obvious, it would be a rare defense lawyer indeed who would gamble his client's fate, and his own license to practice, on a subsequent appeal attacking the bias of a juror, when the lawyer had a chance to remove that very juror with a peremptory challenge. In one of the first federal circuit cases to discuss this aspect of Martinez-Salazar, the Seventh Circuit predicted that "prudent defense counsel will continue to use peremptory challenges to protect their clients against potentially biased jurors, rather than gambling everything on their ability to show bias after-the-fact and to obtain a reversal of a conviction on this basis." (62) The lawyer's choice in this situation is not "a hard choice" at all; it is a very easy choice. Competent representation demands that the defense lawyer correct the trial judge's error by removing the biased juror peremptorily, unless of course it is the defense lawyer's intention to inject error into the case. (63)

This brings us to the second, and perhaps even more troubling, ethical aspect of the Court's no-waiver conclusion. It invites defense lawyers, facing overwhelming evidence of their clients' guilt and therefore little reason to believe that even an impartial jury will do anything but convict, to intentionally infect the jury with a biased juror whenever a trial court erroneously fails to remove that biased juror for cause, thus ensuring a reversal on appeal. Accepting such a strange invitation violates ethical duties that even criminal defense lawyers owe to the integrity of the judicial system. (64)

Quite apart from these ethical issues, the majority opinion in Martinez-Salazar leaves the interplay between waiver and error in an eerie state of meaninglessness. If a defendant is required to use a peremptory challenge curatively, then under Ross there is no error. If a defendant is not required to use peremptory challenges curatively, then under Martinez-Salazar there is likewise no error. As discussed below, we think these two tortured syllogisms, when put together, actually mean that errors of this kind are harmless, but that the Court is unwilling to analyze them under that standard for fear of what such an analysis will say about the whole role of peremptory challenges in the truth-finding process. (65)

It appears that the Court's logic has not impressed some state appellate courts that have already rejected the Martinez-Salazar approach and have indicated that they will continue to view the loss of a peremptory challenge as per se reversible error under their state constitutions. (66) Perhaps the split among state courts would have remained in any case, but the Court has provided weak guidance even for federal courts on the proper analysis to be applied when these sorts of jury selection errors occur.

Why did the Ross and Martinez-Salazar Courts struggle so mightily to avoid the harmless error issue? In some ways, the Court's failure to face the issue directly is puzzling. In other ways, it is typical of the ambivalence many English and American courts have expressed about the peremptory challenge for centuries.

IV. THE HISTORICAL RELATIONSHIP BETWEEN CHALLENGES FOR CAUSE AND PEREMPTORY CHALLENGES

The problem of curative peremptory challenges addressed in Martinez-Salazar raises issues that lie at the heart of the relationship between the challenge for cause and the peremptory challenge. These two kinds of challenges compliment one another in a rather complex fashion, and their complimentary nature is the product of a long and often misunderstood history.

A. Challenges for Cause

From the very beginnings of the English jury in the late 1100s and early 1200s, prospective jurors could be challenged for cause, though there were only three recognized challenges: being related to the defendant by blood, marriage or economic interest. (67) There was no generic challenge for lack of impartiality. (68) Indeed, English jurors before the fifteenth century, like ancient jurors in Greece and Rome, did not have to be impartial. These ancient and medieval juries typically were "presentment" juries, whose functions were investigatory and accusatory, not decisional. (69) Once the presentment juries accused a defendant of a crime, they turned him or her over to the ruler for trial and punishment. (70) Presentment jurors were often selected precisely because they may have had personal knowledge of the alleged crime, or, more commonly, personal knowledge of the accused or of the alleged victim. Thus, the early English challenge for cause was quite narrow, and bore little resemblance to our modern challenge for cause grounded on a lack of juror impartiality. (71)

As the presentment jury began to evolve into the trial jury, and as jurors were called upon not merely to make accusations but also to determine ultimate guilt, the concept of jurors-as-witnesses began to be replaced with the concept of jurors-as-independent-fact-finders. By the end of the fifteenth century, the notion that jurors had to be impartial was firmly entrenched in English common law. (72) Thus, the challenge for cause necessarily broadened to capture prospective jurors who, for a variety of reasons other than the three ancient challenges for cause, simply could not be impartial.

English challenges for cause also expanded because the English venire grew more and more diverse. Originally, presentment jurors were selected by the King's sheriffs exclusively from the nobility. (73) All jurors had to be freehold males under seventy years old, and their membership in the nobility also meant they typically owned substantial real property. (74) As the Crown's political prosecutions increasingly targeted anti-royalist members of Parliament, Parliament began to react, out of sheer self-preservation, by making some pro-defendant changes to criminal procedures, including the methods of jury summoning and selection. Over time, the King's influence in jury selection waned, and specific levels of property ownership began to replace rigid feudal qualifications. Many more kinds of prospective jurors presented themselves for jury duty, and they brought with them hosts of biases not commonly present when all prospective jurors had been selected by the King exclusively from the relatively homogeneous class of loyal nobles. (75)

The final, and perhaps most significant, pressure toward the expansion of the English challenge for cause was the demise of the English peremptory challenge. Prior to 1305, the King's prosecutors had an unlimited number of peremptory challenges, so their resort to challenges for cause was simply unnecessary. As discussed below, the number of allowable peremptory challenges in England steadily declined over the next 700 years, and by 1989 they were eliminated entirely. (76) The broadening of the English challenge for cause probably had as much to do with the narrowing and eventual elimination of the English peremptory challenge as anything else. There is a deep and unbreakable historical counterpoint between challenges for cause and peremptory challenges.

Although the right to jury was less than comprehensively embraced at the constitutional debates, and even in the text of the Constitution itself, (77) the Sixth Amendment did constitutionalize the long-standing English and colonial common law guaranties of an impartial jury. (78) In fact, at least one draft of the Sixth Amendment expressly provided that a criminal defendant had a right to challenges for cause, (79) although this express reference was later deleted at Madison's urgings, on the ground that challenges for cause are so inherent in the concept of an impartial jury that their express mention was unnecessary. (80)

In 1911, Congress codified the Sixth Amendment notion of the right to an impartial jury by creating a right in defendants (and prosecutors) to challenge jurors for cause. (81)

Because the Sixth Amendment's guarantee of an impartial criminal jury is applicable to the states via the Fourteenth Amendment, (82) its implicit requirement that there be some recognized mechanism--challenges for cause--to effect the guaranty of impartiality is also binding on the states. (83) Thus, every state recognizes a criminal defendant's right, and the prosecutor's right, to challenge jurors for cause. By statute or rule or both, most states recognize two fundamentally different types of challenges for cause: ones in which a described relationship between the juror and some trial participant is irrebuttably presumed to render the juror partial ("principal challenges"); and ones in which the challenger must demonstrate to the trial judge that the prospective juror is in fact partial ("challenges to the favour"). (84)

State principal challenges typically provide that litigants may challenge jurors who are related not just to the defendant but also to the lawyers, (85) are in an employee/employer relationship with the defendant, (86) are in a landlord/tenant relationship with the defendant, (87) are in a debtor/creditor relationship with the defendant, (88) are in any fiduciary relationship with the defendant, (89) were witnesses to any of the alleged events, (90) or were jurors in any previous trial regarding the same allegations. (91)

States that have lists of principal challenges also have a catch-all challenge to the favor. These catch-ails typically mimic the broad impartiality language of the Sixth Amendment by providing that all litigants may challenge for cause any jurors who cannot be "impartial" or "fair" or "unbiased." (92)

Interestingly, the federal statutes and rules, and the statutes and rules of a handful of states, do not contain any laundry list of principal challenges for cause. (93) In theory, all challenges for cause in federal courts and in these few state courts are challenges to the favor, and the challenger must actually prove to the judge that the prospective juror, for whatever reason, cannot be fair and impartial. In practice, however, a common law has evolved that mimics the principal challenges. Beginning as early as 1807, state and federal courts have recognized that certain relationships between prospective jurors and litigants give rise to "implied bias," and should be disqualifying even in the absence of satisfactory evidence of actual bias. (94) The laundry list of implied biases includes the same kinds of juror-litigant relationships recognized in most states as principal challenges, including kinship, financial interest and former jury service in the same cause. (95)

B. Peremptory Challenges

Peremptory challenges first appeared in England sometime between 1250 and 1300. (96) They were initially recognized only in criminal cases, and indeed only in capital criminal cases. (97) Only the King's prosecutors, not the defendant, had a right to exercise peremptory challenges, and there was no limit to the number of peremptory challenges the King's prosecutors could exercise. (98)

As large presentment juries began to evolve into smaller trial juries, the peremptory challenge probably began simply as a mechanism to reduce the unwieldy size of the presentment jury. Faced with the prospect of having to cut down large presentment juries to twelve-man trial juries, prosecutors were required by sheer arithmetic to excuse many presentment jurors. (99)

There is some historical support for the proposition that these early peremptory challenges, both to the presentment juries themselves and later to the separately summoned trial juries, were not really peremptory challenges at all but rather were challenges for cause made by an infallible King. If the King's prosecutors asserted that a prospective juror had some connection to the defendant sufficient to require his removal under the then narrow English challenges for cause, that royal assertion was deemed infallible, and no further inquiry was necessary. (100) Some commentators have also suggested that early English peremptory challenges were actually a kind of shorthand challenge for cause in small English villages and towns, where it was commonplace for lawyers, judges, jurors and defendants to be well acquainted with one another and, thus, for cause disqualifictions to be obvious to all. (101)

In an almost immediate response to the King's unlimited peremptory challenges in capital cases, English courts began to permit criminal defendants in capital cases to exercise their own peremptory challenges. By 1300, it was well-settled as a matter of common law that in all capital cases the Crown had an unlimited number of peremptory challenges and the defendant had thirty-five. (102)

In 1305, in a further effort to limit the King's abusive prosecutorial powers, Parliament passed the Ordinance for Inquests, which abolished all prosecutorial peremptory challenges but retained the defendant's thirty-five peremptory challenges. (103) The King's courts immediately side-stepped the Ordinance for Inquests by adopting a common law procedure they called "standing aside," in which the King's prosecutors could direct any number of prospective jurors to "stand aside" pending selection from the balance of the reporting panel. (104) Subject only to the problem of running out of regular jurors and being forced to use some of the jurors who were standing aside, which circumstance was presumably quite rare given the limited nature of challenges for cause, (105) this procedure was of course tantamount to giving the prosecution an unlimited number of peremptory challenges.

From 1305 forward, the number of allowable defense peremptory challenges in English criminal trials steadily decreased. Parliament reduced them from thirty-five to twenty in 1530, (106) to seven in 1948, (107) to three in 1977, (108) and abolished them entirely in 1989. (109) Despite even these reduced numbers, one of the most striking things about the 700-year history of English peremptory challenges is that defendants almost never exercised them. Indeed, they were so rarely exercised that some scholars have posited that ordinary criminal defendants simply had no "right" to any peremptory challenges, despite the pronouncements of Parliament. (110) And although the prosecution theoretically retained its power to ask prospective jurors to stand aside, the standing aside procedure was just as rare, and perhaps even rarer, than the exercise of defense peremptories. (111)

Most American colonial courts accepted the English common law practice of giving criminal defendants some peremptory challenges, but they were in sharp disagreement over the question of whether the prosecution should have any peremptory challenges at all, by standing aside or otherwise. Some colonies allowed prosecutors an unlimited number of peremptory challenges, while others allowed none. (112)

This was the mixed state of colonial affairs at the time of the constitutional debates. There is no record of any discussion by the framers, formal or informal, about peremptory challenges. Article III, Section 2 of the Constitution secures the right to jury in all federal criminal trials, but makes no mention of the peremptory challenge. (113) Indeed, in 1919 the Court held in Stilson v. United States (114) that "[t]here is nothing in the Constitution of the United States which requires the Congress to grant peremptory challenges to defendants in criminal cases; trial by an impartial jury is all that is secured." (115)

Congress codified the English practice in 1790, by giving thirty-five peremptory challenges to defendants in federal treason cases and twenty in all other federal capital cases. (116) It did not address the question of defense peremptories in non-capital federal cases, or the question of whether the prosecution had any peremptory challenges in any kind of federal criminal case. Federal courts in these early years were sharply divided over the question of whether federal prosecutors had a common law right to peremptory challenges, either traditionally or by standing aside. (117) The Supreme Court settled the matter in 1856, when it held in United States v. Shackleford that neither prosecutorial peremptories nor standing aside were rooted in federal common law, and that federal courts must instead look to state procedures to decide whether federal prosecutors could peremptorily excuse prospective jurors or have them stand aside. (118)

In a direct though somewhat delayed response to Shackleford, Congress established in 1865 that in all non-capital federal felony cases defendants would have ten peremptory challenges and the prosecution two, and that in capital cases the prosecution would have five peremptory challenges but defendants' peremptories would be decreased from thirty-five to twenty. (119) In 1872, Congress increased the number of prosecutorial peremptories in non-capital cases from two to three, and for the first time extended the notion of peremptory challenges to federal civil cases (giving each side three) and to federal misdemeanors (giving each side three). (120) In 1911, Congress changed the peremptory challenge numbers again: six for the prosecution and twenty for the defendant in capital cases; six for the prosecution and ten for the defendant in non-capital felony cases; and three each in misdemeanor and civil cases. (121)

When the Federal Rules of Criminal Procedure were adopted in 1946, Rule 24(b) increased the prosecution's peremptories in capital cases to equal the defendant's at twenty. (122) That 1946 version is the scheme currently in place in the federal courts: each side gets twenty peremptory challenges in capital cases; the prosecution gets six and the defendant ten in non-capital felony cases; and each side gets three in misdemeanor and civil cases. (123)

This federal evolution in peremptory challenges was generally mimicked in the states. By 1790, most states recognized by statute a defendant's right to some peremptory challenges, and most states shared the pre-Shackleford federal view that prosecutors had a common law right to some peremptory challenges. (124) Only fourteen years after Shackleford--by 1870--almost all states had enacted statutes granting both the prosecution and the defense some peremptory challenges. (125) Today, every state recognizes some number of peremptory challenges for both sides in criminal and civil cases, though it is interesting that many states continue to follow the asymmetry of federal Rule 24(b) by giving the prosecution considerably fewer peremptory challenges than the defense. (126)

Although state courts were certainly free, even after Stilson, to hold that there is a state constitutional right to peremptory challenges, we have found no such cases. Similarly, even though state courts were free after Shackleford to ground the peremptory challenge in state common law, we have again found no such cases. Thus, peremptory challenges in all state courts, like those in federal court, are creatures of statute and court rule, and are not compelled either by the Constitution or by the common law. (127)

C. Judicial Review

It is important to recognize that peremptory challenges long antedated the right of a criminal defendant to appeal a conviction, both in England and America. Indeed, there was never an English or federal common law right to a criminal appeal. (128) The founders engaged in some limited debates about the question of whether our Constitution should recognize the right to a criminal appeal, (129) but ultimately decided it should not. (130) In 1881, Congress gave the Supreme Court statutory authority to review federal death penalty convictions by writ of error, (131) but it was not until the Court of Appeals Act of 1891 that federal criminal defendants had a generalized statutory right to a criminal appeal. (132)

The evolution of the right to a criminal appeal in the state courts followed a similar pattern. Most colonial and early state courts followed the English tradition of not recognizing any common law right to a criminal appeal. (133) All fifty states eventually recognized the right to a criminal appeal by rule or statute, and none of these provisions has ever been repealed. (134)

The fact that peremptory challenges were part of the criminal trial process long before there was any statutory or constitutional right to a criminal appeal is important when we consider the problem of curative peremptory challenges. As Justice Scalia observed in his concurrence in Martinez-Salazar, peremptory challenges must have had a fundamentally curative purpose in these early years, when they represented the only way to correct a trial judge's erroneous retention of a biased juror. (135) As he also observed, we distort that historical purpose when we allow a criminal defendant the option of either using a peremptory challenge to remove a biased juror or hoarding the peremptory challenge and appealing the conviction. (136)

V. HARMLESS ERROR, STRUCTURAL ERROR AND PLAIN ERROR

Before we examine the problem of the curative use of peremptory challenges and the effect, if any, that an imbalance in peremptory challenges should have on the integrity of criminal convictions, we pause to consider the general notions of harmless error, structural error and plain error. It is, after all, in the context of these larger doctrines that the jury selection errors addressed in this article must ultimately be analyzed, despite the Court's current and rather curious reluctance to do so. (137)

A. Harmless Error

No trials are perfect, and the idea that not all trial imperfections infect the integrity of the truth-finding process is a central limitation to the judicial review of jury verdicts. Indeed, as modern constitutional law and criminal procedure have vastly complicated the trial machinery, there are more opportunities than ever for various "errors" to creep into the trial process, and a correspondingly unprecedented demand that our appellate courts separate the important errors from the unimportant ones. The devil, of course, is in the details, and in the lines we draw between the kinds of trial errors we will and will not tolerate.

It has always been so. Although many harmless error opponents claim our modern harmless error rule is in derogation of English common law, (138) in fact the earliest English pronouncements on the subject recognized a broad concept of harmless error. As early as 1807, in a criminal case involving the erroneous admission of evidence, the King's Bench declared that there must be a new trial only "if the case without such improper evidence were ... not clearly made out ... and the improper evidence might be supposed to have had an effect on the minds of the jury." (139) The precise test of harmlessness changed from case to case, (140) but it was well-recognized throughout the late 1700s and early 1800s, at least in the King's Courts, that not all trial error required reversal. Legal historians often refer to the harmless error rules announced in these early criminal cases as establishing the "Orthodox English Rule." (141)

The Orthodox English Rule was more forgiving than our modern notions of "harmless beyond a reasonable doubt." Under the Orthodox English Rule, all trial errors were subject to the harmless error inquiry; that is, the Orthodox English Rule had no exception, like our pre-Chapman (142) approach, for errors deemed to be particularly significant, or, as we would say, for errors of constitutional magnitude. (143) Moreover, if an appellate court were "satisfied" that the trial error did not affect the trial outcome, the conviction was affirmed. (144)

Not until the mid-1800s did the Orthodox English Rule begin to give way to a rule of per se reversal. An 1835 civil case decided in the Exchequer Courts, Crease v. Barrett, (145) is generally credited as having originated the so-called Exchequer Rule, under which virtually all trial error was automatically reversible. (146) Its legacy was short-lived. Parliament abrogated the Exchequer Rule with its adoption of the Judicature Act of 1873, under which civil trial errors were presumed harmless unless "some substantial wrong or miscarriage occurred." (147)

The history of harmless error in America mirrored its history in England--both in terms of the common law and in terms of legislative responses to the common law. (148) Because the Orthodox English Rule was well entrenched during the early post-Revolutionary period, most colonial courts and their state court successors accepted it. (149) But the Exchequer Rule eventually found its way into our common law, as it did in England. By 1900, most state and federal appellate courts were of the mind that virtually all trial errors, no matter how trivial, required reversal and a new trial. (150) These cases, just as in England, triggered a quick legislative response.

Beginning as early as 1912, states began to enact legislation overruling the per se cases. (151) All fifty states presently have harmless error statutes or rules. (152) In 1919, Congress did the same for the federal courts, enacting its own harmless error statute. (153) The federal harmless error statute remains in effect today, and indeed was incorporated with slightly different language into the Federal Rules of Criminal Procedure, as Rule 52(a). (154)

Until 1967, the federal harmless error statute and rule, and their state counterparts, were widely assumed to be inapplicable to trial errors that deprived defendants of constitutionally based rights. (155) All this changed in 1967, when the Court, in Chapman v. California, (156) held that even constitutional errors can be harmless if the appellate court is sufficiently confident that the constitutional error did not affect the trial outcome. In particular, the Court in Chapman held that a violation of a defendant's Fifth Amendment rights, in the form of the prosecutor commenting in closing arguments on the defendant's decision not to testify, could be harmless. (157) The Chapman Court also held that the prosecution must "prove beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained." (158)

Chapman marked a sea change in the Court's approach to harmless error. For the first time since the Original English Rule, the truth-finding reliability of the process, and not the process itself, once again became the focus of the harmless error inquiry. (159) Since Chapman, the Court has added one constitutional error after another to the growing list of constitutional errors subject to harmless error review, including: impeaching a defendant with his or her post-Miranda silence; (160) admitting a coerced confession; (161) giving the jury over-broad instructions in the sentencing stage of a capital case; (162) giving the jury an instruction containing an erroneous conclusive presumption; (163) and giving the jury a burden-shifting instruction. (164) Indeed, just three years after Chapman, the Court was already describing the notion of applying harmless error review to constitutional errors as the rule, and per se reversals for constitutional errors as the exception. (165)

But the exceptions have not disappeared entirely. There remains a very small group of errors, which the Court has come to label "structural errors," that continue to merit automatic reversal.

B. Structural Error

In Chapman itself, the Court identified three kinds of constitutional rights--the right to counsel, the right to the suppression of a coerced confession and the right to an unbiased judge--as "so basic to a fair trial" that their denial would not be subject to harmless error analysis and would instead compel automatic reversal. (166) The Chapman Court did not articulate in any more detail whether these three particular types of errors were immune from its sweeping change because they involved particularly important rights that justified protection regardless of their impact on trial outcomes, or rather whether they were simply the kinds of error that always affect trial outcomes.

In Arizona v. Fulminante, (167) the Court not only struck one of Chapman's three untouchable categories of errors that required automatic reversal--the erroneous admission of a coerced confession--it also coined the term "structural" error to describe the dwindling group of errors immune to harmless error analysis. (168) Most importantly, the Fulminante Court made it absolutely clear that structural errors require reversal not because they implicate rights abstractly deserving of some greater degree of protection, but rather because those rights are so bound up with the reliability of the process that we irrebuttably presume their violation had an effect on outcome. (169)

The Fulminante Court added three structural errors to the Chapman list it had just reduced by one: (170) the denial of the right to self-representation; (171) the denial of the right to a public trial; (172) and the discriminatory exclusion of members of a defendant's race from a grand jury. (173) In the ten years since Fulminante, the Court has added only one class of errors to the list of structural errors requiring automatic reversal: the giving of a constitutionally deficient instruction on the standard of proof beyond a reasonable doubt. (174)

Despite a firestorm of criticism, both inside the Court's own ranks (175) and in the academy, (176) there is every indication that Fulminante's reliability-based view of harmless error is here to stay. Indeed, in 1999, the Court reaffirmed its commitment to reliability as the pole star of the harmless error inquiry. In Neder v. United States, (177) the Court held that a trial judge's erroneous failure to submit the materiality element to the jury was not structural error, and was instead subject to harmless error analysis. The Neder Court rejected the defendant's argument that by failing to instruct the jury on an element of the offense, the trial court in effect deprived the defendant of the structural right to have a jury decide his guilt beyond a reasonable doubt. (178)

Now that Chapman has taught us that the inquiry is no longer "constitutional v. non-constitutional errors," what factors drive the decision about whether a particular kind of error should be amenable to harmless error analysis? It seems to us that there are two distinct, though related, principles: (1) is the error the kind of error that will likely affect the reliability of the truth-finding process; and (2) is the truth-finding impact of the error incapable of rational assessment? In accordance with the teachings of Chapman and Fulminante, both of these factors focus on the error's effect on the truth-finding process, not on the abstract rights of the defendant.

Only by asking both of these questions do we get to the heart of the proper harmless error inquiry, and to the correct criteria for structural error. If an error is the kind of error unlikely to affect the reliability of the trial outcome, then we need not ask the second question about the difficulty of ferreting out its actual impact. Such errors should be harmless despite the difficulty of assessing actual impact. Indeed, many abjectly harmless errors are nevertheless resistant to any kind of rational impact assessment.

On the other hand, if the error is the kind of error we think likely to affect the reliability of the trial outcome, we will still subject the error to harmless error analysis if we think its impact is capable of being quantified. Thus, in Fulminante, even though all the Justices agreed that the admission of a coerced confession is the kind of error likely to affect the reliability of the trial outcome, the majority concluded that its impact, though great, could nevertheless be measured against the other evidence in a case.

It is only when the error is likely to affect reliability and has effects incapable of being measured, that we should label the error "structural" and reverse without any harmless error inquiry.

The Court's now long-standing commitment to a reliability-based view of harmless error has important implications for the jury selection errors that are the subject of this article. That commitment to reliability supports our thesis that all of these jury selection errors are harmless because they have, by definition, no qualitative impact on the reliability of convictions, even though we might all agree that the effects of any particular jury selection error are almost always impossible to quantify. (179)

D. Plain Error

Appellate courts use the term "plain error" not to describe errors that are necessarily reversible, but rather to describe errors that an appellate court may choose to review even though the defendant failed to object to the errors at trial. (180) We include this discussion of plain error because it bears on the Court's "no waiver" prong of Martinez-Salazar. (181)

Well settled principles of waiver, applied to the trial process, have yielded the common law rule that appellate courts will not generally consider trial errors if a party failed to object during the trial to the procedure or ruling giving rise to the claimed error. This rule--often called the "contemporaneous objection rule"--is not only grounded in basic principles of waiver, it also reflects the sound institutional policy that trial courts should be given an opportunity to correct their own errors by having those errors pointed out by counsel. The contemporaneous objection rule is codified in Federal Rule of Criminal Procedure 52(b), Federal Rule of Evidence 104, and in similar rules in virtually every state. (182)

But the contemporaneous objection rule was never absolute. As early as 1896, the Court recognized that if certain "plain error was committed in a matter so absolutely vital to defendants," then the error could be reviewed on appeal even though it had not been raised in the trial court. (183) By focusing on errors "vital to defendants," the Court's early plain error cases, like its later harmless error cases, (184) appeared to be grounded on reliability: certain errors will be reviewed, notwithstanding the lack of a trial objection, if those errors are likely to have had an adverse impact on a defendant, that is, to have contributed to the conviction. (185) Indeed, Rule 52(b), which codified the common law of plain error, described those errors that are reviewable despite the lack of a trial objection as not only being "plain errors" but also "defects affecting substantial rights." (186) The phrase "substantial rights" is identical to the phrase used in Rule 52(a) to describe harmless error, reinforcing the notion that although these two inquiries are on their face very different and deal with different portions of the process, plain error and structural error may, at bottom, be two ways at looking at the same fundamental problem: when will we reverse convictions without examining the effect the error had on the trial outcome? (187)

The plain error doctrine has undergone an evolution that in many ways mimics the evolution of harmless error. Just as the Orthodox English Rule focused the harmless error inquiry on reliability, so too did the early plain error cases. (188) Just as the Exchequer Rule moved the harmless error inquiry away from reliability and toward the protection of a defendant's rights for the sake of those rights, cases beginning in the 1930s did the same for plain error. (189) Just as Fulminante has taken the harmless error inquiry full circle back to reliability, so too have the Court's plain error cases. (190) Indeed, in United States v. Young, the Court stated that the doctrine of plain error must be used "sparingly," because any unwarranted expansion "would skew [Rule 52(b)'s] `careful balancing of our need to encourage all trial participants to seek a fair and accurate trial the first time around against our insistence that obvious injustice be promptly redressed.'" (191)

In United States v. Olano, the Court completed this evolution back to reliability by articulating a three-part test for plain error: 1) there must be error; 2) the error must be "plain" (that is, "obvious" or "clear"); and 3) the error must "affect substantial rights." (192) The Court described the third prong as requiring the defendant to prove that the error "affected the outcome of the district court proceedings." (193) We cannot imagine a clearer articulation of the reliability standard. Errors are now "plain," just as errors are now "structural," not because they trampled on a defendant's important rights, but rather because the trampling affected the reliability of the conviction.

Currently, the small list of plain errors expressly approved by the Court includes defense counsel's failure to make a motion for judgment of acquittal at the close of the prosecution's case, (194) failure to object to a jury instruction that erroneously requires the jury to presume malice, (195) and the failure to object to confusing sentencing instructions in a death penalty case. (196) The much larger list of errors the Court has held are not plain errors includes defense counsel's failure to object to improper prosecutorial closing arguments, (197) failure to object to the trial court's order allowing two alternate jurors to be present during jury deliberations, (198) failure to object to jury instructions in a perjury case that left out the element of materiality, (199) and the failure to object to the exclusion of the defendant from an in camera heating during trial. (200)

As we discuss below, there is a long tradition of enforcing the contemporaneous objection rule when it comes to jury selection errors; that is, jury selection errors are rarely deemed to be plain errors. (201) Because the plain error inquiry is once again an inquiry based on the reliability of the process, these holdings are entirely consistent with our contention that the sorts of jury selection errors analyzed in this article seldom have any adverse impact on trial outcomes. Thus, they are not only harmless but they should never even be reviewed if a defendant fails to object to them or, more particularly in the case of curative peremptory challenges, if a defendant fails to cure a cause error with an available peremptory challenge.

VI. THE PROBLEM OF THE CURATIVE PEREMPTORY CHALLENGE

Courts have had trouble with the problem of the curative peremptory challenge because they have had trouble articulating the purpose of the peremptory challenge and identifying its role in securing an impartial jury. It is a decidedly odd problem. Defendants are constitutionally entitled to an impartial jury, but they are not constitutionally entitled to peremptory challenges. The peremptory challenge is, in the words of the Martinez-Salazar Court, "auxiliary" to a defendant's right to an impartial jury. (202) So exactly when does a denial of this auxiliary rule-based or statute-based "right" to peremptory challenges rise to the level of a constitutional violation, and a reversible violation at that?

As we suggest below, the judicial ambivalence on this issue reflects a deeper and unarticulated ambivalence about peremptory challenges themselves. (203) On the one hand, if peremptory challenges have any constitutional significance at all, it must be because they serve as a check on the trial court's erroneous rulings on challenges for cause, and therefore act as a screen to increase the chances that biased jurors will not sit. And yet the Court in Martinez-Salazar holds that a defendant is not required under the federal rules to use peremptory challenges curatively. On the other hand, if we concede that peremptory challenges really have very little to do with the selection of an impartial jury, then errors that result in a compromise of a defendant's "right" to peremptory challenges will almost always be harmless. As discussed below, we believe the Court's most recent decisions, first in Ross and now in Martinez-Salazar, portend an acknowledgment that peremptory challenge errors are by their very nature harmless precisely because they do not have, and have never had, much to do with selecting impartial jurors. (204)

In addition to this underlying doctrinal ambivalence about the role of the peremptory challenge, there is a more mundane, but just as important, factor that has confused the law in this area. The problem of the curative peremptory challenge can come up, and has come up, in a variety of different circumstances. Courts have not been terribly good at recognizing, let alone distinguishing, these different circumstances.

We suggest there are four variables to consider: (1) whether the error at issue was the erroneous granting of a prosecution challenge for cause or the erroneous denial of a defense challenge for cause; (205) (2) whether, in the former case, the prosecution exhausted its peremptory challenges; (3) whether, in the latter case, the defendant exhausted his or her peremptory challenges; (206) and (4) whether, again in the latter case, the defendant peremptorily excused the very juror who should have been excused for cause. These four variables in turn generate the following six scenarios:

   Scenario 1: The trial court erroneously grants a prosecution challenge for
   cause, and the prosecution exhausts its peremptory challenges.

   Scenario 2: The trial court erroneously grants a prosecution challenge for
   cause, but the prosecution does not exhaust its peremptory challenges.

   Scenario 3: The trial court erroneously denies a defense challenge for
   cause, the defendant exhausts all his or her peremptory challenges, and the
   defendant uses one of the peremptory challenges on the erroneously retained
   juror.

   Scenario 4: The trial court erroneously denies a defense challenge for
   cause, the defendant exhausts his or her peremptory challenges, but the
   defendant does not use one of the peremptory challenges on the erroneously
   retained juror.

   Scenario 5: The trial court erroneously denies a defense challenge for
   cause, the defendant does not exhaust his or her peremptory challenges, but
   the defendant does use one of the peremptory challenges on the erroneously
   retained juror.

   Scenario 6: The trial court erroneously denies a defense challenge for
   cause, the defendant does not exhaust his or her peremptory challenges, and
   the defendant does not use one of the peremptory challenges on the
   erroneously retained juror.

As we examine the issues of harmless error, structural error, waiver, and the relationship between challenges for cause, peremptory challenges, juror impartiality and due process, we believe it is important to distinguish these six scenarios.

A. AN ANALYSIS OF SCENARIOS 1, 2, 3 AND 5 AFTER MARTINEZ-SALAZAR: HARMLESS ERROR, STRUCTURAL ERROR OR NO ERROR AT ALL?

Let us first address the Court's fiction that these kinds of errors are not errors at all. There is some potency to the idea that an error, if cured, is no longer error. But, of course, that is true only if the cure itself does not inflict another kind of error. In each of these four scenarios the cause error is cured, but the price is the loss, or relative loss, of a peremptory challenge.

In Scenario 3, surely the defendant has been "deprived" of a peremptory challenge when he or she is forced to use a peremptory challenge on a prospective juror that the trial judge should have removed for cause. The palpable error is the denial of the challenge for cause; the immediate effect is that the defense has one fewer peremptory challenge than it otherwise should. This error is as much of an "error," and the deprivation of the peremptory challenge as much of a "deprivation," as when a trial judge simply decides to give the defendant one less peremptory than the rules or statutes provide. (207)

If Scenario 3 error is "error," then Scenario 1 error is also "error," because the effect of both "errors" is an imbalance in peremptory challenges to the defendant's detriment. It is the relative number of peremptory challenges that is important in gauging a litigant's power to shape the jury. A defendant who retains a measly six peremptory challenges in the face of a trial judge's wholly unfounded and erroneous decision to give the prosecution 100 peremptory challenges has surely suffered "error."

We concede that the question of whether Scenario 2 and 5 error is really "error" is a marginally more complex inquiry. The Scenario 2 defendant must contend with the argument that the prosecution did not really get an "extra" peremptory challenge because it did not exhaust its peremptory challenges. Another way to state this argument is that even had the trial court not erroneously granted the prosecution's challenge for cause, the prosecution could still have removed that juror with its unused peremptory challenge. This scenario is Gray v. Mississippi, (208) and the Court expressly, and we think quite correctly, rejected the prosecution's "unexercised peremptory" argument. (209)

Scenario 5 error is the counterpoint to Scenario 3. There is no doubt that the Scenario 5 defendant was deprived of a peremptory challenge by being forced to use one on the erroneously retained juror. However, the argument is that such a defendant is hardly in a position to complain about the loss because he or she did not choose to exhaust the remaining peremptory challenges. This is an argument about the effect of the error, not about whether there is reviewable error in the first instance.

Notwithstanding that Scenario 1, 2, 3 and 5 errors are in fact "errors," we nevertheless contend that the general rule in each of these four scenarios should be that the jury selection error is harmless and that the convictions must be affirmed. (210) In each of these scenarios, no demonstrably biased prospective jurors end up sitting on the jury, so none of the scenarios involves any violation of a defendant's right to an impartial jury under the Sixth Amendment. In other words, all the jurors who actually sit in a Scenario 1, 2, 3 or 5 case were properly passed for cause.

The only effect of a Scenario 1, 2, 3 or 5 error is an imbalance in peremptory challenges. (211) Is this really the kind of error, and really the kind of "right," that justifies reversing otherwise perfectly valid convictions returned by perfectly impartial jurors? The answer must be no. This error is not constitutional, is not structural and is harmless by any measure of that inquiry.

Scenario 1, 2, 3, and 5 error is certainly not constitutional error. Because a criminal defendant has no right to any peremptory challenges, it is difficult to understand how an imbalance in peremptory challenges could rise to the level of constitutional error. Legislatures and supreme courts could abolish all peremptory challenges tomorrow without inflicting constitutional injury on criminal defendants. They could probably even eliminate all defense peremptory challenges and retain all prosecution peremptory challenges. (212) A trial court's unintentional elimination of a single defense peremptory challenge likewise inflicts no constitutional injury.

Defendants nevertheless incant the mantra of due process in these cases, as they often do when an error that has no palpable constitutional reverberations is nevertheless generically claimed to have deprived them of the minimum process required for a fair (that is, reliable) trial. But these incantations ring particularly hollow in Scenarios 1, 2, 3 and 5. We cannot imagine an error with less due process implications than one that results in a defendant still being tried by an impartial jury, in a trial containing all the procedural and evidentiary protections upon which our system is built. (213)

Moreover, since Chapman and Fulminante, it is also clear that Scenario 1, 2, 3 and 5 errors are not structural errors and are therefore subject to harmless error review, even if by some stretch of the imagination we might label them "constitutional" errors. An error is no longer structural just because it is constitutional error or because it involves some other especially important right, or even just because its impacts may be difficult to gauge. (214) The inquiry is reliability. How can it be said that the reliability of a trial is likely to be compromised when a defendant loses a single peremptory challenge, but when all the jurors who actually hear the case are fair and impartial? It cannot. By any sensible measure of "structural error," and certainly by the Court's increasingly strict measures, jury selection error of this sort is not "structural."

This brings us to the harmless error inquiry. If ever there were a category of errors that borders on "harmless as a matter of law," it is the errors in Scenarios 1, 2, 3 and 5. The juries in these Scenarios were vetted for cause, and all of them were, by definition, fair and impartial. Yet they returned convictions. By any measure of harmlessness, depriving a defendant of a single peremptory challenge will surely be harmless in most if not all cases. (215)

The uniqueness of jury selection errors is that by their very nature they offer no resistance to the evidentiary or instructional counterweights we typically place on the scales of harmlessness. Ordinarily, the harmless error inquiry requires an appellate court to ask whether the same jury would have reached the same result with different evidence or different instructions. However, under these scenarios, an impartial jury has already determined, based on all the evidence and the instructions of law, that the prosecution has proved defendant's guilt beyond a reasonable doubt. The jury selection errors in Scenarios 1, 2, 3 and 5 require appellate courts to ask whether a different but equally impartial jury would have reached the same result with the same evidence and the same instructions. The very foundations upon which our system is built--that cases are decided based on the law and the evidence, and not on the peccadilloes of the fact-finders--will almost always require us to conclude that such errors are harmless.

B. AN ANALYSIS OF SCENARIOS 4 AND 6 AFTER MARTINEZ-SALAZAR: WAIVER, INVITED ERROR, AND THE COURT'S STRANGE INVITATION

Even though Martinez-Salazar was not a Scenario 4 or 6 case, the majority went considerably out of its way, as discussed above, to hold that Federal Rule 24(b) does not require a defendant to use peremptory challenges curatively. (216) Under this view, Scenario 4 and 6 defendants can never be said to have waived the challenge for cause argument by failing to cure the claimed error with a peremptory challenge. As a result, what had once been relatively well-settled rules of waiver in Scenarios 4 and 6 are now in great doubt, at least in the federal courts.

We contend that the Court in Martinez-Salazar was wrong in its "no waiver" holding, and that under the most basic principles of waiver Scenario 4 and 6 defendants have waived the cause errors by failing to take advantage of the opportunity to cure the errors with available peremptory challenges. (217) This position is not only compelled by the doctrines of waiver, plain error and invited error, but the contrary position encourages defense lawyers to engage in the very kinds of games playing that puts the trial system in public disrepute.

Scenario 4 and 6 defendants take Justice Ginsburg's advice and elect not to peremptorily excuse a palpably biased juror. Instead, they keep that biased juror on the jury, save the peremptory challenge for use on other prospective jurors, then try to obtain a reversal on appeal on the ground that the juror was biased. The only difference between Scenario 4 and 6 is that in the former the defendant ends up exhausting all peremptory challenges and in the latter he or she does not. By making this "hard choice," the Scenario 4 and 6 defendants suffer no imbalance in peremptory challenges. The issue is a much more direct and traditional one: does the defendant waive the cause error on appeal because he or she elected not to peremptorily excuse the very juror whom the defendant later claims on appeal to have been biased? Just asking the question makes the answer evident, but let us analyze these two waiver scenarios a little more comprehensively.

We begin with the well-established principle that cause errors themselves are not plain errors. That is, if a defense lawyer fails to challenge a juror for cause during jury selection, the cause error is ordinarily waived on appeal. (218) If a lawyer waives a cause error by not making it during jury selection, we cannot understand how that same lawyer preserves the error by making the challenge for cause, having it erroneously denied, and then failing to remove the juror with an available peremptory challenge. In both cases, the lawyer had the power to remove the biased juror, elected not to do so, and now wants to complain about that election on appeal. (219)

The salutary purposes of the contemporaneous objection rule are particularly important in jury selection. We want to encourage the trial court's opportunities to correct jury selection error for the very reason that we are at the beginning of the whole process. There are obvious and enormous judicial economies in correcting jury selection errors when they are made, rather than with an appeal and re-trial. Moreover, the challenge for cause and Batson inquiries are intensely factual. The last kind of system any of us should want is one in which appellate courts are either forced to be the initial finders of fact on these issues or forced to reverse convictions precisely because they do not want to be the initial fact finders. The trial court judge is in a unique, indeed exclusive, position to make the factual calls upon which these assertions of jury selection errors depend. (220) Finally, although challenges for cause and Batson challenges are not always easy decisions for trial judges, once a decision favorable to the challenger is made the corrective action is easy. Challenges for cause are granted and the challenged jurors removed; Baston challenges are sustained and either the wrongfully excused juror returned to the jury or a new venire brought in. (221)

The point is that there are many powerful institutional reasons for having these issues raised early, raised before the judge who is in the best position to rule on them, and corrected early. The same arguments apply to Scenario 4 and 6 errors. The defendant is in a position to correct the trial judge's cause error early and easily, simply by using an available peremptory challenge to strike the problem juror. That corrective action is far superior, by any legitimate institutional measure, to letting the biased juror sit through an entire trial and then reversing the conviction and ordering a new trial.

Even without analogy to these two other kinds of non-plain jury selection errors, Scenario 4 and 6 error is precisely the kind of error that has no bearing on reliability and thus could never join the increasingly exclusive club of plain errors. Indeed, these kinds of errors are not only not plain errors, they are a special category of self-induced errors--often called "invited errors"--that appellate courts have traditionally been loath to entertain. Some of the most obvious examples of invited errors include: a party offering evidence at trial and having it admitted, then objecting to its admission on appeal; (222) a party successfully keeping out evidence at trial by way of a motion in limine, then objecting on appeal to its exclusion; (223) a party successfully resisting the other side's motion to recuse the trial judge, then arguing on appeal that the judge was biased; (224) a defendant successfully requesting a particular jury instruction, then objecting to the giving of that instruction on appeal; (225) a defendant stipulating to an element of the offense (a prior conviction), then objecting on appeal that the trial court instructed the jury that that element had been stipulated. (226) Less obvious examples include a defendant "opening the door" to otherwise inadmissible evidence or improper argument. (227)

Scenario 4 and 6 errors are the kind of invited jury selection errors that appellate courts should not review. Defendants in these scenarios had a straightforward and easily exercised opportunity to prevent biased jurors from sitting on their juries, yet they chose not to take that opportunity. Their choice may have been a "hard choice," but it should not be without consequences. (228) By electing not to remove biased jurors peremptorily, defendants participate in their seating every bit as much as the trial judges who erroneously fail to remove them for cause. Appellate courts should not entertain these self-induced errors.

VII. RETHINKING THE CONNECTIONS BETWEEN PEREMPTORY CHALLENGES AND RELIABILITY

The problem of curative peremptory challenges forces us to come to grips with a fundamental question: what is the purpose of peremptory challenges? The manner in which the Martinez-Salazar Court resolved the problem suggests the Court is on the brink of abandoning the myth that peremptory challenges help to achieve impartial juries. That abandonment could be the long-predicted beginning of the end of peremptory challenges.

Some judges and commentators have argued that peremptory challenges are unconstitutional because they violate equal protection, (229) due process (230) and the Sixth Amendment's guarantee of a fair cross-section. (231) The counter-arguments have always centered on the value of peremptory challenges as a tool to insure impartial juries. (232) If in fact peremptory challenges have little or nothing to do with the goal of selecting impartial jurors, which is what we think Martinez-Salazar may actually mean, then they will not likely survive constitutional attack. (233)

The role of peremptory challenges, and in particular their connection with the goal of selecting impartial jurors, has been rather unclear ever since juror impartiality became fixed in the common law in the fifteenth century. (234) Indeed, the very fact that peremptory challenges antedated the notion of juror impartiality by some 200 years should tell us something about whether juror impartiality and peremptory challenges have much to do with one another. (235) Blackstone's famous and oft-quoted description of the peremptory challenge as "a provision full of that tenderness and humanity to prisoners, for which our English laws are justly famous," (236) was coupled with a preceding, less famous and seldom quoted description of it "as arbitrary and capricious." (237)

We inherited this English ambivalence about peremptory challenges, as well as the English proclivity only to articulate their sunny side. Ever since it acknowledged in Stilson that peremptory challenges are not constitutionally required, (238) the Court has consistently spoken about them in reverential, almost constitutional, terms. The irony about these descriptions in dicta is that they have always been coupled with non-dicta limitations on peremptory challenges, either on their discriminatory exercise or on the effect they are deemed to have on trial outcomes.

Thus, in Swain v. Alabama, the majority describes the peremptory challenge as having "very old credentials" and as "a necessary part of trial by jury," (239) but then holds for the first time that peremptory challenges are not entirely peremptory, and that the Equal Protection Clause prohibits prosecutors from systematically using them in a racially discriminatory fashion. In Batson v. Kentucky, the Court repeats Swain's refrain in extolling the virtues of the peremptory challenge, (240) but then holds that peremptory challenges cannot be used discriminatorily in the single case at hand, whether or not the prosecutor has a history of systematically using them discriminatorily. In all the post-Batson cases, the Court continues to pay homage to the role of peremptory challenges, but then continues to extend Batson's encroachment into the "peremptoriness" of peremptories. (241) In Ross v. Oklahoma, the Court again describes peremptory challenges as a way to ensure an impartial jury, yet holds that their impairment is not only not reversible error but not error at all.

It is not until Martinez-Salazar that the Court begins to talk about peremptory challenges with language that suggests they may not be as important to the re-emerging paradigm of reliability as earlier cases had assumed:

   We have long recognized the role of the peremptory challenge in reinforcing
   a defendant's right to a trial by an impartial jury. But we have long
   recognized, as well, that such challenges are auxiliary; unlike the right
   to an impartial jury guaranteed by the Sixth Amendment, peremptory
   challenges are not of federal constitutional dimension. (242)

Perhaps we read too much into the tea leaves of the Court's language, but we detect a distinct and rather remarkable change in the way the Court has expressed its long-standing ambiguity about the peremptory challenge. What in the days of Swain and Batson was described as an important tool to ensure jury impartiality, in whose constriction the Court only reluctantly participated to protect more important rights of equal protection, is in the post-Martinez-Salazar world an anachronism of common law whose abject violation is not even recognized as error.

VIII. CONCLUSION

Jury selection errors whose only impact is a net reduction in peremptory challenges are harmless. When a criminal defendant uses a peremptory challenge to remove a prospective juror whom the trial court should have removed for cause, the defendant is using peremptory challenges in precisely the curative manner for which they were intended. If the defendant in such circumstances is convicted, he or she is convicted by a fair and impartial, albeit different, jury, and the conviction should be affirmed. (243) If the defendant in such circumstances elects not to expend a peremptory challenge on the erroneously retained juror, then that defendant should not be allowed to raise the cause error on appeal.

Martinez-Salazar comes as close as the Court has ever come to talking about peremptory challenges in the language of harmless error. We hope it will continue to do so, and that it will continue its commitment to reliability as the pole star of harmlessness. We hope the Court will re-think its "no waiver" holding, and return to well-settled principles of waiver and non-plain error that are entirely consistent with the reliability paradigm. Whatever it does in the future, the Court seems finally to be talking about peremptory challenges in a more realistic, less romanticized, way. That is good news for a system that needs less games playing in jury selection and more common sense on appeal.

William T. Pizzi, B.A. Holy Cross; M.A. University of Massachusetts; J.D. Harvard. Professor of Law and Byron White Center Fellow, University of Colorado School of Law. I would like to thank my research assistant, Matthew D. Grove, who was tremendously helpful on the research in this article.

Morris B. Hoffman, B.A. University of Colorado; J.D. University of Colorado School of Law. District Judge, State of Colorado, Second Judicial District (Denver). I would like to thank my law clerks, Catherine A. Woods and Christian H. Hendrickson, for their invaluable contributions to some of the research in this article. The views I express here are of course my own, and do not necessarily reflect the views of the District Court for the Second Judicial District or any of my colleagues on that Court.

(1.) Because virtually all criminal cases are tried to juries, virtually all criminal cases that result in convictions have the potential to generate a jury selection issue on appeal. Moreover, the rate at which criminal convictions are appealed is roughly 100%. See Steven Shavell, The Appeals Process as a Means of Error Correction, 24 J. LEGIS. STUD. 379, 421 n.80 (1995). The ubiquity of jury selection issues in appeals is thus not surprising. See JOY CHAPPER & ROGER HANSON, UNDERSTANDING CRIMINAL APPEALS 32 (1989) (finding jury selection errors the eighth most common error in criminal appeals).

(2.) State v. Doleszny, 508 A.2d 693 (Vt. 1986).

(3.) Id. at 694.

(4.) At least there is nothing in the opinion of the Vermont Supreme Court indicating the juror was rehabilitated. Cf. State v. Grega, 721 A.2d 445, 481 (Vt. 1992) (noting rehabilitation took place after juror disclosed she knew and respected physician-witness and would be upset if someone said he did anything wrong, yet later said she would not hold such criticism against defendant or prosecution and could listen to both sides).

(5.) In the introductory portions of this article, we use the phrase "automatic reversible error" to describe an error that demands reversal of a criminal conviction without any inquiry into the actual impact of the error on the reliability of the truth-finding process. As described in greater detail in infra text accompanying notes 139-80, courts and commentators have used several synonyms to describe this sort of error, including "per se reversible error" and, most recently, "structural error."

(6.) See e.g., State v. Sexton, 787 P.2d 1097, 1099 (Ariz. 1989) (finding defendant's right to peremptory challenges "so substantial that forcing a party to use a peremptory challenge to strike potential jurors who should have been stricken for cause denies the litigant a substantial right"); People v. Macrander, 828 P. 2d 234, 236 (Colo. 1992) (finding prejudicial error occurred where trial court failed to strike a potential juror who was related to an attorney of record and defendant was forced to use a peremptory challenge to remove the juror); People v. Scott, 566 N.Y.S.2d 399, 400 (N.Y. App. Div. 1991) (finding new trial necessary where the trial court erred in failing to remove a juror for cause and defendant exhausted his peremptory challenges). For the federal courts, the Supreme Court may have resolved the issue in United States v. Martinez-Salazar, 528 U.S. 304 (2000), which we discuss extensively later in this article. See infra notes 49-65 and accompanying text. Before Martinez-Salazar, several federal circuits held that jury selection error resulting in a defendant's loss of a peremptory challenge was automatically reversible. See, e.g., United States v. Hall, 152 F.3d 381, 408 (5th Cir. 1998) ("`The denial or impairment of the right to exercise peremptory challenges is reversible error without a showing of prejudice.'") (citation omitted); United States v. Cambara, 902 F.2d 144, 148 (1st Cir. 1990) ("[I]t is true that restricting a defendant's use of the lawful number of peremptory challenges is reversible error if a challenge for cause is erroneously denied.").

(7.) See, e.g., Pickens v. State, 783 S.W.2d 341, 345 (Ark. 1990) (finding loss of peremptory challenges not to be reversible error); People v. Pride, 833 P.2d 643, 663 (Cal. 1992) (finding no basis for reversal where defendant fails to prove the trial court erroneously denied any challenges for cause); State v. Pelletier, 552 A.2d 805, 810 (Conn. 1989) (finding defendant unharmed because he was not forced to accept any juror or alternate whom he requested to be removed for cause); State v. Barlow, 541 N.W.2d 309, 311-13 (Minn. 1995) ("[T]he necessity to exercise a peremptory challenge to strike a juror whom the trial court has erroneously refused to remove for cause does not deprive the defendant of a fair trial."); State v. Lindell, 629 N.W.2d 223, 251 (Wis. 2001) overruling State v. Ramos, 564 N.W.2d 328 (Wis. 1997) ("The substantial rights of a party are not affected or impaired when a defendant chooses to exercise a single peremptory strike to correct a circuit court error."). In the federal courts, there were several circuits, even before Martinez-Salazar, that applied a harmless error analysis to these kinds of errors. See, e.g., United States v. Brooks, 161 F.3d 1240, 1245-46 (10th Cir. 1998) (finding trial court's erroneous failure to remove a juror for cause harmless error where defendant has not alleged that any of the jurors actually seated were biased); United States v. Farmer, 923 F.2d 1557, 1566 (11th Cir. 1995) (rejecting "`the notion that the loss of a peremptory challenge constitutes a violation of the constitutional right to an impartial jury'") (citation omitted).

(8.) State v. Ramos, 808 P.2d 1313 (Idaho 1991).

(9.) Actually, three prospective jurors initially expressed the belief that Mr. Ramos must have done something by virtue of the fact he was on trial. Id. at 1314. Two were excused on defense challenges for cause, but the trial judge denied the challenge for cause as to the third. Id.

(10.) Or more precisely, there was no indication in the opinion of any rehabilitation.

(11.) 808 P.2d. at 1314.

(12.) As the Idaho Supreme Court put it:

   Ramos made no showing in his motion for new trial that he was prejudiced by
   being required to use a peremptory challenge to remove [the juror]. Ramos
   has not demonstrated, nor has he even suggested, that any of the other
   jurors remaining on the panel were not impartial or were biased. Thus, if
   there was any error, it was harmless.

Id. at 1315.

(13.) 528 U.S. 304 (2000).

(14.) See infra notes 205-06 and accompanying text. In addition to errors in rulings on challenges for cause, there is of course a second basic kind of jury selection error--error in a trial court's application of Batson v. Kentucky, 476 U.S. 79 (1986). The Supreme Court has never squarely addressed the issue, but it has suggested in various cases, including Batson itself, that Batson error is per se reversible. See, e.g., Vasquez v. Hillery, 474 U.S. 254, 263-64 (1986); Batson, 476 U.S. at 100 ("If the trial court decides that the facts establish prima facie discrimination and the prosecutor does not come forward with a neutral explanation of his actions, our precedents require that Petitioner's conviction be reversed."). Perhaps because of these strong signals from the Supreme Court, lower federal courts have uniformly held that Batson error is per se reversible. See, e.g., United States v. McFerron, 163 F.3d 952, 956 (6th Cir. 1998); United States v. Hall, 152 F.3d 381, 408 (5th Cir. 1998); Tankleff v. Senkowski, 135 F.3d 235, 238 (2d Cir. 1998); United States v. Underwood, 122 F.3d 389, 392 (7th Cir. 1997); United States v. Annigoni, 96 F.3d 1132, 1143 (9th Cir. 1996); Ford v. Norris, 67 F.3d 162, 170 (8th Cir. 1995); Ramseur v. Beyer, 983 F.2d 1215, 1225-26 (3d Cir. 1992).

The state courts, however, are split on this issue. Compare K.S. v. Carr, 618 So.2d 707, 712 (Ala. 1993) and State v. Pharris, 846 P.2d 454, 459 (Utah Ct. App. 1993) (Batson error reversible per se) with Seubert v. State, 749 S.W.2d 585, 588 (Tex. Crim. App. 1988) (Batson subject to harmless error analysis), rev'd on other grounds, 787 S.W.2d 688 (Tex. 1990).

Professor Muller has suggested on the one hand that the Supreme Court's hints about this issue are far from obvious--for example, in Allen v. Hardy, 478 U.S. 255 (1986), the Court held that Batson is not retroactive, in part because it only has marginal beating on the truth-finding function in individual cases, compared to its primary purpose of strengthening public confidence in the system. Eric L. Muller, Solving the Batson Paradox: Harmless Error, Jury Representation, and the Sixth Amendment, 106 YALE L.J. 93, 121-24 (1996). On the other hand, Professor Muller concludes that the Sixth Amendment implications of a Batson error justify classifying it as per se reversible. Id. at 97-107, 149-50; see also Susan N. Herman, Why the Court Loves Batson: Representation-Reinforcement, Colorblindness, and the Jury, 67 TUL. L. REV. 1807 (1993) (arguing that Batson error is not properly amenable to harmless error analysis because Batson implicates a bundle of rights that go far beyond the reliability of the outcome in a particular case).

Whether Batson error should be per se reversible or should be subject to harmless error analysis, or indeed whether Batson itself should be reconsidered, is beyond the scope of this article. But it is important to recognize that the adoption of a Batson per se rule does not compel a per se rule for the jury selection errors we discuss in this article. Even if we concede both Professor Muller's point--that Batson in fact protects a defendant's case-specific right to an impartial jury and therefore has everything to do with the reliability of the case outcome--and Professor Herman's point--that Batson implicates a wide range of institutional issues going well beyond the reliability of a particular case--neither observation applies to errors that simply result in a small imbalance of peremptory challenges. This is so, as discussed below in infra text accompanying notes 217-28, because the loss of a peremptory challenge, as opposed, arguably, to the discriminatory use of peremptory challenges, has little to do with the impartiality of juries and therefore little to do with the reliability of trial outcomes.

(15.) See infra notes 210-33 and accompanying text. Although the problem of the curative use of peremptory challenges has been the subject of many judicial opinions in state and federal courts, we have found only one law review article written on the topic: William G. Childs, The Intersection of Peremptory Challenges, Challenges for Cause, and Harmless Error, 27 AM. J. CRIM. L. 49 (1999). Mr. Childs' article was written while the Court's decision in Martinez-Salazar was pending, and he urged the Court to hold that the impairment of peremptory challenges is a per se violation of a criminal defendant's rights to due process, at least in federal courts. Id. at 52, 75-79. Of course, we disagree with Mr. Childs' analysis, and believe that he seriously overvalues the historical and practical importance of peremptory challenges, undervalues their mischievousness, and ignores the Court's unmistakable movement toward reliability as the measure of harmlessness.

This article is limited to a discussion of the problem of curative peremptory challenges in criminal cases. But a similar problem exists in civil cases. Civil litigants, like criminal defendants, are entitled to impartial jurors, to peremptory challenges, to the protections of Batson and in general to have their civil jury trials conducted in accordance with the minimum standards of fairness imbedded in the due process clause. They are also subject to the doctrines of harmless error and waiver. Indeed, there are a handful of reported cases discussing the problem of curative peremptory challenges in a civil context, and reaching conflicting conclusions about whether there is reversible error. Compare Kirk v. Raymark Indus., Inc., 61 F.3d 147, 162 (3rd Cir. 1995) (reversing civil judgment when litigant forced to expend peremptory challenge to cure trial judge's error in denying challenge for cause) with Getter v. Wal-Mart Stores, Inc., 66 F.3d 1119, 1123 (10th Cir. 1995) (affirming civil judgment despite fact that litigant forced to expend peremptory challenge to cure trial judge's error in denying challenge for cause). We limit our inquiry in this article to the criminal side because the problem of the curative use of peremptory challenges in a criminal case is in many ways more difficult than in a civil case. The notion of harmless error in civil cases is much broader--that is, more forgiving--than in criminal cases, and more "process" is generally "due" to a criminal defendant than a civil litigant. As a result, the problem of curative peremptory challenges in a criminal case presents significantly starker issues about the role of peremptory challenges and the extent to which their impairment justifies reversing trial outcomes. If, as we argue in this article, criminal convictions should not be reversed when a defendant's peremptory challenges are impaired by a cause error, then it necessarily follows that civil judgments should not be reversed in similar circumstances.

(16.) 481 U.S. 648 (1987).

(17.) Although the record in Gray was not terribly clear, what was clear is that the beleaguered trial judge committed a cacophony of jury selection errors at almost every opportunity. During the death qualification of the jury, four or five of the panel members expressed doubts about whether they could ever impose the death penalty, and the prosecution challenged these prospective jurors for cause. Id. at 652-54. Under Wainwright v. Witt, 391 U.S. 510 (1968), and Witherspoon v. Illinois, 469 U.S. 412 (1985), the trial judge should have inquired further of those jurors to determine whether they should in fact have been removed for cause, although there are also hints in the record that some of them may have been feigning concerns about the death penalty to avoid jury service. Gray, 481 U.S. at 652-53. In any event, neither counsel nor the trial judge pressed these jurors on their views about the death penalty. Instead, the prosecutor simply used peremptory challenges to remove them. A new prospective juror was then called into the jury box, whom the prosecutor ultimately decided to strike peremptorily, but by then, the prosecutor had exhausted his peremptory challenges. The trial judge seemed to accept some responsibility for the situation, suggesting that he had not excused the earlier prospective jurors for cause because he had wrongly assumed that the prosecutor had plenty of peremptory challenges left. Id. In an apparent effort to undo the damage of his earlier failures, the judge struck the juror in question for cause, even though her answers were not a sufficient basis for a challenge for cause under Witherspoon-Witt. Id. at 654-55.

(18.) The trial judge's performance in Gray is a good advertisement for the superiority of the so-called "strike" method of exercising peremptory challenges over the so-called "sequential" or "jury box" method. Under the "strike" method, prospective jurors are called to an expanded box equal in number to the number of all regular jurors, alternates and total peremptory challenges. These prospective jurors are then examined, and challenges for cause are made and ruled upon. If a challenged juror is removed for cause, a replacement is called up to the expanded box, questioned and challenged for cause if appropriate. Only after all the jurors in the expanded box have survived challenges for cause do the parties then exercise their peremptory challenges. By contrast, under the "sequential" or "jury box" method, only enough jurors are called to the box to accommodate the number of regular jurors and alternates. These jurors are then questioned, challenged for cause and then peremptorily challenged, juror-by-juror. That is, if a prospective juror survives challenges for cause, the parties must then decide whether to accept that prospective juror or exercise a peremptory challenge. For a good description of these two methods see Brent J. Gurney, The Case for Abolishing Peremptory Challenges in Criminal Trials, 21 HARV. CRIM. & CONST. L. REV. 227, 244-56 (1986) and Leonard B. Sand, Batson and Jury Selection Revisited, 22 LITIG. 3, 3-4 (1996). The errors in Gray, of course, were profoundly compounded because the trial judge used the traditional "sequential" method for peremptory challenges, hopelessly and seamlessly interlacing his decisions on challenges for cause with the parties' exercise of peremptory challenges.

(19.) Gray v. State, 472 So.2d 409, 423 (Miss. 1985). The Mississippi Supreme Court's decision was even less clear than the trial court record. The United States Supreme Court was forced to engage in the rather remarkable exercise of guessing the basis for the Mississippi Supreme Court's decision. The Court divined that the Mississippi Supreme Court's opinion was susceptible to three very different interpretations, and proceeded to address each of them. 481 U.S. at 661-62.

(20.) 481 U.S. at 661-62.

(21.) Id.

(22.) Id. at 668.

(23.) Id. at 665.

(24.) Id. at 672-80 (Scalia, J., dissenting, joined by Rehnquist, C.J., and White and O'Connor, JJ.).

(25.) 487 U.S. 81 (1988).

(26.) Id. at 83-84

(27.) Most courts applied Witherspoon-Witt symmetrically, to disqualify not only inappropriately adamant anti-death penalty jurors but also inappropriately adamant pro-death penalty jurors. The Court formally adopted this view in Morgan v. Illinois, 504 U.S. 719 (1992).

(28.) Under Oklahoma law, a capital defendant like Mr. Ross was entitled to nine peremptory challenges. OKLA. STAT. tit. 22, [section] 655 (1981).

(29.) 487 U.S. at 87-88.

(30.) Id.

(31.) See infra note 222.

(32.) See 487 U.S. at 89-90 (citing McDonald v. State, 15 P.2d 1092, 1094 (Okla. 1937); Stott v. State, 538 P.2d 1061, 1064-65 (Okla. Crim. App. 1975); and Ferrell v. State, 475 P.2d 825, 828 (Okla. Crim. App. 1970)).

(33.) 487 U.S. at 90-91.

(34.) Id. at 91.

(35.) Id. at 92-98 (Marshall, J., dissenting, joined by Brennan, Blackmun and Stevens, JJ.).

(36.) Id. at 91 n.4.

(37.) See, e.g., Karen T. Grisez, Ross v. Oklahoma: A Reversal of the Reversible Error Standard in Death-Qualification Cases, 38 CATH. U. L. REV. 881 (1989) (analyzing both cases as death penalty cases and not as applying to jury selection errors generally).

(38.) 528 U.S. 304 (2000).

(39.) From what we can glean from the Ninth Circuit opinion, 146 F.3d 653 (9th Cir. 1999), the case appears to have been a rather routine drug case with very strong evidence against both defendants. Other than the jury selection issue that went to the Supreme Court, the only other issue discussed in the Ninth Circuit opinion was Mr. Martinez-Salazar's contention that there was not enough evidence to convict him of the gun charge. The Ninth Circuit summarily rejected this contention because the gun in question was found under the front passenger seat next to the heroin, Mr. Martinez-Salazar had been seated in that seat, and Mr. Martinez-Salazar admitted to a drug agent that the gun was always in the car. Id. at 654.

(40.) Id. at 655.

(41.) The transcript of the inquiry into the juror's bias reads as follows:

      THE COURT: On your questionnaire, you said in question number eight, the
   answer: "I would favor the prosecution." Is that--are you saying that you
   would not be able to listen to the evidence, and decide what happened, and
   follow the instructions of the Court, but would simply vote for a
   conviction because people are charged with drug crimes?

      JUROR: No. I think what I'm saying is all things being equal, I would
   probably tend to favor the prosecution.

      THE COURT: You understand that one of the things the jury will be told,
   of course, is that the prosecution, the Government has the burden of
   proving someone guilty beyond a reasonable doubt. And I suppose
   realistically, all things being equal wouldn't be beyond a reasonable
   doubt. Would you disagree with that?

      JUROR: No, I guess I wouldn't disagree with that.

      THE COURT: I guess the important question is--and perhaps let me ask it
   this way. It's kind of my question. But if you were the defendants here
   charged with this crime, and all of the jurors on your case had your
   background and your opinions, do you think you'd get a fair trial?

      JUROR: I think that's a difficult question. I don't think I know the
answer to that.

      Martinez-Salazar's trial counsel, Mr. Garcia, then followed up by
questioning the juror:

      MR. GARCIA: If you were to error [sic], where would you feel more
   comfortable erring, in favor of the prosecutor or the defendant?

      JUROR: Well, again, not having heard any evidence in the case, I think
   that's kind of hard to say. I think, as I indicated on here, I would
   probably be more favorable to the prosecution. I suppose most people are. I
   mean they're predisposed. You assume that people are on trial because they
   did something wrong.

      THE COURT: Well, you see, you heard me out there when I started the
   trial. That's not the general proposition. If it is, it's wrong. It's
   contrary to our whole system of justice. When people are accused of a
   crime, there's no presumption--

      JUROR: There's a--

      THE COURT:--of guilty. The presumption is the other way. That's the way
our system--

      JUROR: I understand that in theory.

      THE COURT: Okay, all right, all right. Why don't you wait, and we'll be
   done here in a few minutes, okay? Thank you very much.

Id.

(42.) Id.

(43.) The Ninth Circuit concluded that trial judge erred when he denied the defendants' challenge of the subject juror for cause:

   The district court here should have excused [the juror] for cause because
   he did not and would not affirmatively state that he could lay aside his
   admitted bias in favor of the prosecution. [The juror] clearly acknowledged
   this bias, even after being instructed by the district court that it was
   "contrary to our whole system of justice." He never retreated from his
   statement of bias.

Id. at 656. In its arguments in the Supreme Court, the government did not contest the Ninth Circuit's conclusion that the trial court erred in denying the defendants' challenge for cause to this juror. 528 U.S. at 309.

(44.) Under Federal Rule of Criminal Procedure 24(b), a defendant charged with a non-capital felony is entitled to ten peremptory challenges and the prosecution six. FED. R. GRIM. P. 24(b); see infra notes 122-23 and accompanying text. The rule explicitly states that co-defendants must share these ten peremptory challenges, but also provides that in the case of multiple defendants the trial court "may" allow additional peremptory challenges and/or direct that they be exercised separately. Trial judges have great discretion in deciding whether to award co-defendants additional peremptory challenges. See, e.g., United States v. Magana, 118 F.3d 1173, 1206 (7th Cir. 1997); State v. Cambara, 902 F.2d 144, 147-48 (1st Cir. 1990). Moreover, Rule 24(c) provides that a defendant shall have an additional peremptory challenge if up to two alternate jurors are to be seated. FED. R. CRIM. P. 24(c).

In Martinez-Salazar, one alternate juror was seated, so the two co-defendants had a presumptive eleven peremptory challenges between them. Moreover, the