Kathleen Hackett, a consumer of bread purchased at retail, filed a complaint under Section 4 of the Clayton Act, 15 U.S.C. Sec. 15 (1971) seeking treble damages, costs and attorney's fees from the defendants for their alleged violation of Sections 1 and 2 of the Sherman Antitrust Act, 15 U.S.C. Secs. 1 and 2 respectively. The defendants are seven bakers of pan-baked bread who sell such bread through various channels of distribution in the Philadelphia market area. Mrs. Hackett's complaint seeks recovery for injury to herself and to all other members of a class whom she claims to represent, consisting of all individual consumers in the Philadelphia market area who have purchased pan-baked bread from retail stores for their own consumption or for consumption by members of their family. The parties agree that her individual claim is for roughly nine dollars. She asserts that some one and one-half million purchasers of bread from retail stores, who purchased the bread involved for the use of some six million consumers comprise the class, of which she claims to be an appropriate class representative.

Mrs. Hackett's complaint was inspired by an indictment, United States of America v. General Host Corp., Crim. No. 23,200 (E.D.Pa., filed Mar. 13, 1968), which charged that the defendants conspired to fix the prices and terms of sales of bread in the Philadelphia market. The district court accepted a plea of nolo contendere to that indictment. There is pending in the district court a civil action Philadelphia v. General Host Corp., Civil No. 68-704 (E.D.Pa., filed Apr. 2, 1968) in which Philadelphia, on behalf of itself and various other public and private institutions, is seeking recovery of the damages of such institutional purchasers occasioned by the alleged price fixing conspiracy.

The defendants moved for a pretrial conference and a stay of all proceedings until the district court should, pursuant to E.D.Pa. R 45(c), determine whether the case should be maintained as a class action. See Fed.R.Civ.P. 23(c) (1). By an order dated May 4, 1970 and in accordance with Rule 23(c) (1), the district court fixed a schedule for the filing of affidavits and briefs dealing with the class action issue. On July 30, 1970, the court filed an opinion and order, the crucial provisions of which follow:

"It is our considered opinion that under the facts of this case, the problem of management of the proposed class is clearly insurmountable. The class is so large that it would be unmanageable and could only result in knotty, complicated and unnecessary problems.

The plaintiff's request for confirmation of her action as a class action is Denied."

Mrs. Hackett then requested the district court to amend the July 30, 1970, order to set forth a statement, pursuant to 28 U.S.C. Sec. 1292(b) (1971) that its decision on the class action issue involves a controlling question of law as to which there is substantial ground for difference of opinion and that immediate appeal would materially advance the ultimate termination of the litigation. On August 25, 1970 the district court declined to issue such Sec. 1292(b) certification.

Mrs. Hackett then filed a notice of appeal. She contends that the July 30, 1970, order is a final appealable order within the meaning of 28 U.S.C. Sec. 1291. The defendants contend that the appeal is interlocutory and should be dismissed.

No request was made to the district court that the July 30, 1970, order be treated for purposes of Fed.R.Civ.P. 54 (b), as a dismissal of the complaint on behalf of the one and one-half million class members or for a determination and direction pursuant to that rule, that the order be entered as a final judgment. Cf. Hayes v. Sealtest Foods Division of National Dairy Products Corp., 396 F.2d 448 (3d Cir. 1968).

Mrs. Hackett predicates her claim that this court must entertain her appeal upon the so called "death knell" rule of the Second Circuit which first appeared in Eisen v. Carlisle & Jacquelin, 370 F.2d 119 (2d Cir. 1966), cert. denied, 386 U.S. 1035, 87 S. Ct. 1487, 18 L. Ed. 2d 598 (1967). In that case Mr. Eisen sued both for himself and on behalf of all odd lot purchasers and sellers on the stock exchange. His individual claim amounted to only seventy dollars, and in a previous appeal in Eisen Judge Kaufman wrote:

". . . We can safely assume that no lawyer of competence is going to undertake this complex and costly case to recover $70 for Mr. Eisen.

******

* * *

Dismissal of the class action in the present case, however, will irreparably harm Eisen and all others similarly situated, for, as we have already noted, it will for all practical purposes terminate this litigation. Where the effect of a district court's order, if not reviewed, is the death knell of the action, review should be allowed."

The court in Eisen relied upon Cohen v. Beneficial Industrial Loan Corp., 337 U.S. 541, 69 S. Ct. 1221, 93 L. Ed. 1528 (1949), as authority for its "death knell" rationale. In this circuit we have referred to the Cohen rule as the "collateral order doctrine," and we have not been overly hospitable to requests for its extension. See e. g., Borden Co. v. Sylk, 410 F.2d 843 (3d Cir. 1969); but cf. Greene v. Singer Co., Civil No. 71-1835 (3d Cir., filed Nov. 2, 1971). We have not, however, ever been confronted directly with the problem of the appealability of a determination adverse to the confirmation of a class which leaves a single plaintiff to litigate a small claim for money damages.

Indeed no other circuit has had occasion either to adopt or expressly to reject the "death knell" rational.1 The experience of the Second Circuit has shown the application of the rule to be somewhat difficult. In Green v. Wolfe Corp., 406 F.2d 291 (2d Cir. 1968), cert. denied, Trostee, Singer & Co. v. Green, 395 U.S. 977, 89 S. Ct. 2131, 23 L. Ed. 2d 766 (1969), Green had an individual claim amounting to less than $1000 for overcharges on publically traded securities. The court held that an order dismissing the class action aspects of the complaint was appealable. In City of New York v. International Pipe and Ceramics Corp., 410 F.2d 295 (2d Cir. 1969) and Caceres v. International Air Transport Association, 422 F.2d 141 (2d Cir. 1970) where the individual claims of the plaintiffs seeking to represent a class were $1,560,000 and $150,000, respectively, the court declined to apply the Eisen rule and dismissed the appeal. Most recently, in Korn v. Franchard Corp., 443 F.2d 1301 (2d Cir. 1971), the court considered two appeals. In one the appellant Korn, whose individual claims totaled $386, appealed from an order revoking a class action designation. In the other, the appellant Millberg, and her husband, who was acting as her attorney, had claims totaling $8,500. Millberg appealed from an order denying class action designation. Korn's appeal was held to fall within the Eisen rule while Millberg's appeal was dismissed. These results disturbed Judge Friendly, who in concurring wrote:

"Since I regard Judge Feinberg's opinion as correctly applying the present law in this circuit, I concur therein. However, despite the obvious appeal of the 'death-knell' doctrine, I am not sure it affords a rule that is truly workable or, indeed, is legally sustainable. If my fears should be realized, I might wish on some subsequent occasion to request that the court consider in banc whether we are not obliged to formulate a rule that will avoid the necessity of making such ad hoc judgments as have been required in these and other cases and also will afford equality of treatment as between plaintiffs and defendants. Perhaps, before occasion for doing this should arise, we shall have received enlightenment from the Supreme Court." 443 F.2d 1301 at 1307.

Judge Friendly's mention of equality between plaintiffs and defendants refers, of course, to the peculiarity of the "death knell" rationale that it operates only in favor of the plaintiff who has unsuccessfully sought to be designated as a class representative. It neither requires nor permits general supervision by the court of appeals over class action designations.2 Indeed the rule operates in a very narrow area.

First, Eisen does not operate at all in those cases in which federal jurisdiction depends upon jurisdictional amount. In Snyder v. Harris, 394 U.S. 332, 89 S. Ct. 1053, 22 L. Ed. 2d 319 (1968) the Supreme Court rejected the contention that the revised Fed.R.Civ.P. 23 permitted aggregation of claims for purposes of jurisdictional amount. Since the court which developed the Eisen "death knell" doctrine has in Korn v. Franchard Corp., supra, drawn the survivability line as low as $8,500, it is clear that in no case where a party with a $10,000 claim is before the court will the "death knell" doctrine be applicable. In the absence of a party's meeting the dollar amount requirements, under Snyder v. Harris, there will be no federal jurisdiction unless the claim falls within a category of federal question cases which may proceed without regard to such jurisdictional amount. See Zahn v. International Paper Co., 53 F.R.D. 430 (D.Vt.1971).

Secondly, Eisen is not needed to afford interlocutory appellate review in those cases in which the refusal to grant class action designation amounts to a denial of a preliminary injunction broader than would be appropriate for individual relief. 28 U.S.C. Sec. 1292(a) (1). See e. g., Oatis v. Crown Zellerbach Corp., 398 F.2d 496 (5th Cir. 1968); Shapiro, Bernstein & Co. v. Continental Record Co., 386 F.2d 426 (2d Cir. 1967); Brunson v. Board of Trustees, 311 F.2d 107 (4th Cir. 1962). This category of interlocutory appeals is adequate, we think, to protect against most district court inhospitability to class action litigation involving civil rights, the elective franchise, protection of the environment and the like. The Eisen rule is not likely to be needed in this area. Indeed the underlying assumption of Eisen-that no competent counsel would be willing to represent an individual small-claimant-is, whatever its validity, completely inapplicable to those numerous areas where the protection of rights incapable of measure in money is the primary object of the lawsuit.

Thus the "death knell" rule would operate only in that narrow category of cases where the object of the suit is the recovery of money damages, and where a statute affords federal jurisdiction regardless of amount. This narrow category is, however, significant for consumer advocates. But even in this field, the "death knell" rationale, based upon the assumption that no competent lawyer would undertake a complicated case to recover a small amount of money, must be qualified by several considerations. A number of federal statutes provide for the award of counsel fees and costs.3 Indeed the statute under which Mrs. Hackett sues contains such a provision. 15 U.S.C. Sec. 15 (1971). If the "death knell" is to ring for her case the rope is in her attorney's hand. There is no certain basis for the assumption that an interested holder of a small claim will be unable to prevail upon a private attorney to pursue that claim in the hope of being compensated by the award of "reasonable counsel fees" against a wrongdoer. Moreover the burgeoning in recent years of interest in publicly supported legal service organizations and of private support for legal aid and public interest law firms cannot be disregarded.4 Many small but important claims heretofore, for purposes of litigation, beyond the pale of financial practicability, have been successfully litigated by such organizations in recent years. Thus an adverse class action decision may ring out as a death knell on far fewer occasions than superficial analysis would suggest.

This brings us precisely to the practicabilities of the operation of the "death knell" rule. It will operate primarily if not exclusively in that class of cases in which attorneys are willing to undertake on a contingent fee basis class actions for the recovery of money damages for claimed violations of federal regulatory statutes. These are chiefly the federal antitrust5 and securities6 statutes.

The chief policy argument in favor of a hospitable attitude toward such class actions is that they tend to reenforce the regulatory scheme by providing an additional deterrent beyond that afforded either by public enforcement or by single-party private enforcement. Viewed in this light the revised Rule 23 may be seen as an extension by the Supreme Court, acquiesced in by Congress, of the deterrent policies of such statutes as Sec. 4 of the Clayton Act.7 Even assuming such a deterrent policy, however, it must be balanced against the competing policies which have historically protected the federal appellate courts from being overwhelmed by interlocutory appeals.8 And on the side of the balance opposed to interlocutory appellate review as of right must be weighed the existence and effectiveness of alternative discretionary appellate remedies with which to protect against excessive inhospitability of district courts to consumer class actions.

One of these is certification pursuant to 28 U.S.C. Sec. 1292(b) (1971). Another is an order certifying a case as final under Rule 54(b). It has been suggested that an order may not fall under both Sec. 1292(b) and Rule 54(b). C. Wright, The Interlocutory Appeal Act of 1958, 23 F.R.D. 199, 203 (1959). Whether this is true with respect to class action determinations we need not here decide. Cf. Hayes v. Sealtest Foods, supra.9 It suffices to say that by one route or the other a disappointed applicant for class action designation could, with the cooperation of the district court, in those rare instances where it might be appropriate bring before this court for review the questionable Rule 23 order. We have had no indication that the district courts of this circuit will reject applications under Sec. 1292(b) or Rule 54(b), arbitrarily or in disregard of the policy considerations favoring, where feasible, consumer class actions warranted by federal statutes. If in isolated instances arbitrariness creeps in, there remains the ultimate remedy of mandamus. See Interpace Corp. v. City of Philadelphia, 438 F.2d 401 (3d Cir. 1971).

The advantage of restricting interlocutory appellate review of Rule 23 class action determinations to the Sec. 1292(b), Rule 54(b), or mandamus routes is that in each instance at least one court and in the case of Sec. 1292(b), two courts, will exercise some discretion in the allowance of the appeal and weigh the countervailing considerations. With the exception of mandamus,10 those discretionary routes were not open in 1949 when the Supreme Court decided Cohen v. Beneficial Industrial Loan Corp., supra. Since the Interlocutory Appeals Act of 195811 and the 1963 Amendment to Rule 54 (b)12 there is less reason for enlarging those categories of orders which for purposes of appealability are deemed to be sufficiently "collateral" as to qualify under Sec. 1291.

Nor may our decision in Greene v. Singer Co., supra be considered as reflecting any basic change in our attitude toward enlargement of the Cohen doctrine.13 In the Greene case we held that an order denying a defendant's motion to disqualify the plaintiff's attorney on the ground that he had previously represented the defendant in the matter which was the subject of the lawsuit was appealable under Cohen. Even that issue is a close one. See 9 J. Moore, Federal Practice p110.13 at 189 et seq. The considerations favoring "collateral order" treatment in the Greene situation are different and more compelling than here. The decision appealed from in Greene allowed an attorney who had switched allegiance to represent the plaintiff, subject only to a protective order. The attorney's duty of loyalty was the matter in issue on the motion to disqualify, a matter wholly separate from the merits of the lawsuit. Postponement of review of the order denying the disqualification motion until final hearing could have the effect of destroying the very subject matter-the duty of loyalty-which was the object of the collateral proceeding which resulted in the order appealed from.

This case could possibly be viewed as the reverse of an analogy to Greene since the decision refusing confirmation of the proposed class can be construed as one which, in effect, disqualifies Hackett's attorney from representing the one and one-half million other potential claimants.14 Looked at as an order disqualifying an attorney, the order refusing to permit a class action confronts us with some practical problems of standing. The one and one-half million potential claimants are not before the court seeking the attorney's services. This contrasts with the Greene case, where the attorney's former client came before us seeking to preserve his duty of loyalty. Mrs. Hackett's disinclination to proceed with her lawsuit unless her attorney is allowed to represent many others besides herself does not move us to convert by an ipse dixit an order which as to her is clearly interlocutory into a final appealable order. We come down, then, to the question whether on the authority of Greene v. Singer Co., supra, or by a different ipse dixit we should recognize the attorney deprived of the quixotic opportunity15 of representing one and one half million potential claimants, should be recognized as a private attorney general with standing of his own to appeal the adverse class action decision. When all is said and done this pragmatically is the core issue, though conventional pieties about the role of the legal profession might suggest its obfuscation. Realistically, when we are asked to grant interlocutory appellate review of an adverse class action determination we are asked to recognize a separate interest of the attorney sufficient to bring the class action determination within the "collateral order" doctrine, or to recognize the standing of the attorney's client to assert such an interest on his behalf. We decline to do either. Those typical consumer class actions in which the Eisen rule would be likely to operate involve areas of federal law in which public enforcement co-exists with private remedies. There is no compelling need to go beyond those inducements to the bar which already encourage a lively pursuit of private enforcement remedies.

One part of Judge Rosenn's dissenting opinion requires special comment, since it points up a difference in approach to the proper role in society of the limited resources of the judicial process. He says that despite the statute providing for attorneys' fees in antitrust suits he:

". . . cannot believe that a capable attorney would press forward a complex anti-trust action when the amount in controversy is nine dollars . . . Nor does the growth of public legal services, also cited by the majority, provide the means by which an individual litigant might secure legal representation. No publically supported legal service organization could explain or justify the time and effort required to press a nine dollar anti-trust suit." (Dissenting opinion at 631) (footnotes omitted).

Judge Rosenn may well be right on both points. But, if so, the conclusion that the judicial process must therefore provide a mechanism, by making class action determinations appealable, whereby the lawsuit will be more attractive to attorneys, does not follow. If the public interest issue involved in the individual suit is so insignificant that neither a private nor a public attorney deems it worthy of pursuit, despite the availability of an award of attorneys' fees in the event of success, then the public interest issue may well be so insignificant that the redress of the nine-dollar wrong should from a policy viewpoint be left to the realm of private ordering. Our scarce federal judicial resources cannot be allocated on the assumption that they must provide a forum for the vindication of every individual wrong however slight. As we pointed out hereinabove, the only instances in which the adverse class action determination will be made appealable under either the Eisen rationale or the "collateral order" rationale, will be those cases in which a federal regulatory statute permits a suit for money damages regardless of jurisdictional amount. In every such instance other regulatory mechanisms are available to assert the public interest against the wrongdoer. If in some cases as Judge Rosenn suggests the individual claim often will be so small that neither private nor public lawyers think it should be litigated, then that decision of the legal marketplace may be the best reflection of a public consciousness that the time of the lawyers and of the court should best be spent elsewhere.

There remains for consideration whether on this record we should treat Mrs. Hackett's notice of appeal and briefs as a petition for mandamus and grant relief under 28 U.S.C. Sec. 1651 (1971). See e. g. In re Harmon, 425 F.2d 916 (1st Cir. 1970); International Products Corp. v. Koons, 325 F.2d 403 (2d Cir. 1963). As in Interpace Corp v. Philadelphia, supra, the district court's decision was within the limits of the discretion allowed by Rule 23(1) (c). The non-arbitrary exercise of that discretion will not be disturbed by the extraordinary remedy of mandamus.

In view of our conclusions on appealability we have no occasion to rule upon Mrs. Hackett's standing as a plaintiff under Hanover Shoes, Inc. v. United Shoe Machinery Corp., 392 U.S. 481, 88 S. Ct. 2224, 20 L. Ed. 2d 1231 (1968). That issue was tendered by the defendants as additional support for the district court's order.

The appeal will be dismissed.

MAX ROSENN, Circuit Judge (dissenting).

Almost 25 years ago, Mr. Justice Jackson was impressed by the formidable volume of judicial writing on the issue of "the finality of a judgment for purposes of appeal." Dickinson v. Petroleum Conversion Corp., 338 U.S. 507, 508, 70 S. Ct. 322, 94 L. Ed. 299 (1949). The passage of almost a quarter of a century and the promulgation of the new class action Rule 23 has not only added to the girth of the judicial writing on the subject but also to its convolutions.

I share with the majority some of their apprehensions for the judicial machinery because of the burdens, pressures and problems generated by the new Federal Class Action Rule. I do not believe, however, that the fear of potential abuses of the new Rule 23 should require us to hold that an order of the district court refusing confirmation of a class shall for all practical purposes be precluded from appellate review. In reaching such a conclusion, I do not pass judgment on the merits of this case nor on the propriety of the action of the district court in finding this class action unmanageable. I do believe, however, that this action is appealable.

It is important to recharacterize the facts before us to show how they relate to the complex and confused question of "finality" under Section 1291. In this case, a collateral order not going to the merits of the anti-trust claim has effectively terminated further consideration of the case on the merits. In reviewing it, there are two considerations to keep in mind: (1) the order under consideration is wholly collateral to the merits; and (2) its effect is to preclude consideration on the merits so that it affects the plaintiff's substantive case in a decisive, practical sense. Each of these considerations goes to separate definitions of "finality," each of which, in my opinion, is a sufficient ground for finding that there is an appealable, final order here.

The Supreme Court in Cohen v. Beneficial Finance Co., 337 U.S. 541, 69 S. Ct. 1221, 93 L. Ed. 1528 (1949), held that a decision of a district court, requiring a plaintiff in a derivative action to comply with the New Jersey statute ordering a posting of a bond to cover expenses, was: "[an] order [that] . . . did not make any step toward final disposition of the merits of the case and [would] not be merged in final judgment. When that time comes, it will be too late effectively to review the present order, and the rights conferred by the statute, if it is applicable, will have been lost, probably irreparably." (Id., at 546, 69 S. Ct. at 1225) As a result, the order was considered "final" under Section 1291. There are essentially three elements that must be satisfied under the Cohen test: (1) the rights determined must be separable from and collateral to the merits of the case; (2) the question must be too important to deny review at this point; and (3) the time for review is ripe and cannot await the ultimate final judgment because the rights may be lost irreparably through the delay. 9 Moore's Federal Practice p 110.10, at 133.

Subsequent cases clarified the scope of the Cohen rule. Swift & Co. Packers v. Compania Colombiana del Caribe, 339 U.S. 684, 70 S. Ct. 861, 94 L. Ed. 1206 (1950), held that the order of the district court vacating the attachment of a maritime lien was appealable under Cohen. The Court stressed two elements: (1) the issue was collateral to the merits; and (2) restoring the right to attachment and the lien at a later time "would be an empty rite." (Id. at 689, 70 S. Ct. 861) Shortly thereafter, Roberts v. United States District Court for Northern Dist. of Cal., 339 U.S. 844, 70 S. Ct. 954, 94 L. Ed. 1326 (1950), held the denial of a petition to proceed in forma pauperis collateral and appealable under Section 1291. These two cases demonstrate that the important rights which may be lost through delay need not rise to the level of the constitutional questions involved in Cohen.

So long as the decision affects the basic right of the litigant to proceed, or finally determines a claim separable from and independent of the substantive issues, so that a later review of it when an appeal is taken from the final judgment on the merits becomes an empty rite, the Cohen test is met. This element of irreparably affecting important rights of a party to the litigation has become a cornerstone of the judicial approach to collateral order questions.

More recently, in Diaz v. Southern Drilling Co., 427 F.2d 1118, 1123 (5th Cir. 1970), the Fifth Circuit noted that one important reason for allowing review of the default judgment against one of the parties in the case was that it affected the litigant's ability to "make important decisions about its further participation in this suit. . . ."

This approach to the appealability of a collateral order under Section 1291 is independent of any discretionary appeal under Section 1292(b).1 If class action orders are collateral, appeal as of right would exist independently and exclusively of the 1958 Act. See Note, Interlocutory Appeal from Orders Striking Class Action Allegations, 70 Colum.L.Rev. 1292, 1302-3 (1970).

I conclude then that when an order will irreparably affect the basic rights of a party and is sufficiently collateral to the merits that subsequent decisions will not involve this question, it has sufficient finality under Cohen to be appealable under Section 1291.

Appellant meets those tests in this case. Her appeal is from a decision refusing to confirm the class she is attempting to represent. The basis for the decision is grounded solely on considerations of the proper administration of F.R.Civ.P. Rule 23. The determination in no way is dependent upon, or affects, rights asserted by the appellant under the Sherman and Clayton anti-trust acts. The district court's opinion that the class is unmanageable in no way resolves the substantive question of whether the appellees engaged in an anti-trust price fixing conspiracy that raised the retail price of bread in the Philadelphia area. The decision on manageability is final. Even if Mrs. Hackett were to proceed individually with her case, as is suggested, nothing more would ever be decided about the manageability of the class she once sought to represent. The record on that question is complete, the issue fully argued, and the decision based on the fullest consideration of the district judge. Review of that determination at this point by this court would not in any way invade the province of the district court in its determination of any of the issues still before it.

Furthermore, the district court's denial of the right to proceed as a class action has certainly affected the subsequent conduct of the litigation. Assuming that the suit will continue on the part of the appellant in her individual capacity, it is inconceivable that either the lawyers or the court will give a $9.00 suit the same consideration and attention due a class action involving hundreds of thousands of persons and many millions of dollars in potential damages. If nothing else, it would be a waste of judicial resources to allocate extensive time to anti-trust litigation over this paltry sum. Even if the appellant had the economic power and her lawyers the fortitude to pursue this complicated anti-trust action for $9.00 against this group of bakeries, there would be a strong impulse on the part of the appellees to settle it and not incur the heavy legal fees generated by such an action.

Should the plaintiff proceed to judgment in her individual capacity, the parties would be bound by it, absent error of law. To the extent that the denial of the class action will affect how the case is litigated, the damage will be irreparable. This court's reversing the denial of the class will not permit the parties to reopen the judgment on the merits. To the extent this possible result compels the defendants to expend large sums defending the action, although it be for only $9.00, they too have been irreparably damaged if on appeal we confirm the denial of the class by the district court.2

(The Dickinson, Gillespie, Eisen Formulation)

The second consideration referred to in determining "finality" of judgment is that the order refusing to confirm the class action effectively terminates the litigation. This definition of finality was first developed in Dickinson v. Petroleum Conversion Corp., supra, in which the Court had to decide whether an appellant had appealed in timely fashion from a decision of the district court. The issue turned on which of two decisions was "the final one" for purposes of the litigation. In the first decision, all of the party's rights were adjudged, but the court reversed decision on apportioning the fund generated by the judgment awarding damages. The appellant did not appeal the first order but waited for the second judgment. The Court balanced "the inconvenience and costs of piecemeal review on the one hand and the danger of denying justice by delay on the other" (Dickinson, 338 U.S. at 511, 70 S.Ct. at 324) and concluded that the appellant should have appealed from the first decision.

The decision did not mention Cohen, although Cohen had been decided only a year earlier. This is understandable because Dickinson dealt with appellate review of the merits of the case, not the collateral issue. The considerations involved in such a decision are different than under the Cohen rule. Finality in Dickinson meant that the merits of the case, insofar as the appealing party is concerned, are effectively determined. The case, as a whole, is at an end for him. Review does not interfere with the district court's jurisdiction because there is, for all intents and purposes, nothing more for the district court to do.

This line of reasoning was reaffirmed in Gillespie v. United States Steel Corp., 379 U.S. 148, 85 S. Ct. 308, 13 L. Ed. 2d 199 (1964), in which the Court held that it had jurisdiction to review the denial of several causes of action under Section 1291. The mother of a dead seaman sued her son's employer under the Jones Act and under general maritime law in a wrongful death action. She joined as coplaintiffs the decedent's dependent brothers and sisters. The district court struck off the general tort claim and held the siblings could not recover under the Jones Act, thereby foreclosing their participation in the action. The Court used the same balance struck in Dickinson: it looked at the inconvenience in piecemeal litigation as balanced against the danger of denying justice. More importantly, it added that the decision on the propriety of the district court's decision was "fundamental to the further conduct of the case," (Gillespie, 379 U.S. at 153, 85 S.Ct. at 311) and held a Section 1291 appeal proper.

Both of these cases illustrate that the Court was using a simple balancing of interests in determining when it should inject itself into appellate review of questions involving the principal matters in the case before the district court. Under this principle, "finality" meant a determination of whether the issue raised at an early stage of the trial was of such character as to effectively dispose of the merits of the case; it did not mean "finality" as used in the collateral rule test under Cohen.

Eisen v. Carlisle & Jacquelin, 370 F.2d 119 (2d Cir. 1966), followed and relied on the Dickinson-Gillespie formulation, but in a matter involving a question collateral to the merits of the case. Judge Kaufman seems to have reasoned that because of the economic considerations involved in the size of appellant's claim, the denial of the class action was practically the final decision in the case, and therefore, it was proper to invoke the balancing of interests test to determine if this point was the proper place for the appeals court to step into the litigation. Piecemeal litigation seemed unlikely and the demands of justice in the case compelled Judge Kaufman to conclude Section 1291 applied.

Subsequently, the Second Circuit commented upon Eisen in Caceres v. International Air Transport Association, 422 F.2d 141 (2d Cir. 1970). In explaining the history of the prior case, it stated that the earlier decision was based upon Cohen and Roberts, even though the court made perfectly clear that the test involved had nothing to do with collateral questions, but with Dickinson-Gillespie's balancing of interests test. Such a reinterpretation of the cases obviously was bound to cause confusion and the ultimate obfuscation of Cohen and the collateral order rule. More importantly, it is clear that by the time the Second Circuit arrived at Caceres it no longer really struck a balance.3 Rather it had converted the Dickinson-Gillespie doctrine into one question: is there enough money in the single claim that the case will proceed? That point is made perfectly clear by Korn v. Franchard Corp., 443 F.2d 1301 (2d Cir. 1971), in which a party with a $386 claim was allowed to appeal a denial of a class designation, but a party with an $8,500 claim was not. One party was expected to be able to proceed; the other was not. (Id. at 1307).

Such an automatic test, so far removed from the original balancing of the inconvenience of piecemeal litigation against the denial of justice, bothered Judge Friendly who concurred with great reservation. (443 F.2d at 1307). I agree that such an arbitrary formulation hardly illuminates the underlying issues. My Brother Gibbons' discussion of whether Mrs. Hackett could continue this litigation in her individual capacity if she wished, illustrates the problems with the Korn test, and how arbitrary that test is. I believe the proper procedure is to return to the principles from which Korn arose.

In fact, under either the Dickinson-Gillespie doctrine or under the simplified test of Korn, Mrs. Hackett has properly brought her appeal before this court. Suffice it to say at this point that under the present Second Circuit Korn test, if Eisen's $70 claim was too small to carry forward his class action, then this $9 claim is also too small.

More importantly, under the Dickinson-Gillespie formulation, if I analyze the balance of potential piecemeal litigation against the denial of justice, I must again conclude that Mrs. Hackett's appeal should be heard now. Although the majority strongly suggests that Mrs. Hackett could continue this case on her own, it is clear that the potential for piecemeal litigation here is small. She is unlikely to continue the litigation and that is all that is important in striking the balance. Moreover, there are strong questions of justice involved. When the 1966 amendments to the Federal Rules were issued, the A.B.A. Special Committee on Federal Rules of Procedure stated that it hoped that there would be adequate appellate review of many of the questions surrounding the amendments to Rule 23 so that the contours of the amendment would be quickly shaped and clarified. 38 F.R.D. 95, 104 (1965). Besides these general societal interests, the individual needs of justice for Mrs. Hackett and her class are best served by immediate appellate review. As I have stated before, the character of a suit for $9 is simply not the same as that for several million dollars. It is in the interest of both the defendants and the plaintiff to know precisely where they stand with regard to the amount being litigated.

Finally, for purposes of properly understanding the distinctions between Cohen and Eisen, it is important to note what the cases did not say. Many of the Cohen cases involve situations which effectively terminate the litigation. For instance, in Roberts, the failure to be granted leave to proceed in forma pauperis was most likely to terminate litigation. However, in holding that the denial was appealable, the Court merely cited Cohen, which made no mention of the possibility that the collateral order would have ended the litigation. That omission is significant because the "importance" criterion established in Cohen left open the possibility of balancing such a factor. In fact, the collateral order in Cohen requiring posting of a bond presumably had precisely that litigation ending effect. (Note, Appealability, 70 Colum.L.Rev. 1292, 1304 (1970)). Although at least one Cohen case has weighed the "importance" question on the basis of its effect in terminating litigation (Redding & Co. v. Russwine Construction Corp., 135 U.S.App.D.C. 153, 417 F.2d 721, 726 (1969)), it should not generally be thought to be necessary to the Cohen test.

Eisen cites Roberts without comment but presumably for the proposition that collateral decisions that end litigation are appealable. However, its primary test is the one developed in Dickinson and Gillespie and there is never any mention made of the actual Cohen rule that all important collateral matters irreparably affecting the litigation are appealable.

In any case, under either the collateral order doctrine announced in Cohen or the termination of litigation doctrine of Dickinson-Gillespie, I am compelled to find that the Supreme Court mandates the decision denying confirmation of the class be held appealable.

The majority opinion suggests, in effect, that the refusal of the district court to confirm the class action does not necessarily terminate the plaintiff's suit; that although the class action may not be maintained, the suit may be pressed on an individual basis by Mrs. Hackett and others similarly situated. Reference is made to several federal statutes which provide for an award of reasonable attorneys' fees should Mrs. Hackett, or others who follow, desire to press forward.

Indeed, the statute under which Mrs. Hackett proceeded not only provides for reasonable attorneys' fees but also provides that one may sue "without regard to the amount in controversy." Nevertheless, I cannot believe that a capable attorney would press forward a complex antitrust action when the amount in controversy is nine dollars.4 If attorneys could be found willing to go forward on an individual basis with Mrs. Hackett's claim, and the claims of others who allegedly comprise the aggrieved class, dockets would indeed be intolerably overcrowded. Nor does the growth of public legal services, also cited by the majority, provide the means by which an individual litigant might secure legal representation. No publicly supported legal service organization could explain or justify the time and effort required to press a nine dollar anti-trust suit. Moreover, publicly supported legal service organizations generally may not proceed on a contingent fee basis.5

The majority suggests that instead of allowing interlocutory appeal as of right we weigh the effectiveness of "alternative discretionary appellate remedies" for consumer class actions. As I have already indicated, the appeal before us is not interlocutory. Further, "the alternative discretionary appellate remedies" suggested offer little consolation to Mrs. Hackett since she has been denied by the district court the opportunity to use them. Nor do I see any necessity for restricting her to mandamus for appellate review. It has been sparingly applied for such purposes and I am not convinced that it is a reasonable alternative for even an interlocutory appeal. Moreover the variety and number of alternative remedies leaves no assurance that denial of appeal under Sec. 1291 will serve the majority's purpose to protect "the federal appellate works from being overwhelmed by interlocutory appeals".

On the assumption that this is an interlocutory appeal and not an appeal from a final order, the majority expresses the fear that this court may be plagued by interlocutory review. Until we have further experience with class actions, it seems to me that this fear is premature. As the majority points out, its decision operates within a very limited sphere. Indeed, as they say, "no other circuit has had occasion to either adopt or expressly reject the 'death knell' rationale." Moreover, anti-trust litigation seems to be growing slowly in comparison with other fields. For example, in terms of the civil cases commenced in the United States District Courts, civil rights showed an increase of 28.9% in fiscal 1971 over fiscal 1970, commerce cases a 77.9% increase, and securities, commodities and exchange cases a 62% increase,6 while anti-trust cases rose by a mere 15%.7 Further, if it is true that there are "alternative discretionary remedies" which may bring to us the issues in cases similar to this one, it would appear that denying review in this particular case, while allowing review in others, does nothing except to arbitrarily shut the door to this one.

Some comment is in order on our recent decision in Greene v. The Singer Co., (Civil No. 71-1835, 3d Cir., Filed Nov. 2, 1971). Of course, I agree with the majority that the Greene case differs from the case at bar in terms of the factual situation presented. However, I cannot agree that the rule which dictated our conclusion in Greene would warrant a different decision in the case now before us. Greene was, in fact, a reaffirmance of the rules which were first articulated in Cohen, and which have been discussed above. As we said in Greene: "To require appellant to await a final judgment on the merits before testing the legality of the order denying the disqualification may, for practical purposes, deny it the reality of appellate processes." (Slip opinion p. 3). Similarly, in the case now before us, as I have already pointed out, denying appeal in this case will undoubtedly "for practical purposes, deny [Mrs. Hackett] the reality of appellate processes."

It should also be noted that the rationale which motivated the Greene case has been impaired by the decision which the majority takes today. The court in Greene did not decide that a failure to allow appeal would deny appellant the "reality of appellate process." It found only that it might deny appellant that reality. It did not hold that if the appeal did not go forward, appellant would suffer irreparable injury; it held only that there existed "the possibility of irreparable injury." (Emphasis supplied) Mrs. Hackett has at least shown this much.

One final point deserves attention. While the small size of these claims, considered individually, means that for all practical purposes this law suit ends here, the small size does not mean that these claims are insignificant. In fact, this case may illustrate the usefulness of a class action in the litigation of small consumer claims.8 Bread is an essential household staple which is purchased regardless of the economic level of an individual family. Those whose incomes are meager feel the pinch of price fixing in such an area. Yet they are the ones who most need to combine resources in order to pursue their rights in the courts.

I would hold therefore that the order of the district court entered July 30, 1970 comes within the collateral rule of Cohen, supra, or in the alternative is final under Dickinson-Gillespie, and, therefore is an appealable order.