Sufficient factual background for an understanding of the rulings we are about to make in this extraordinary “class action” is to be found in our first opinion Eisen v. Carlisle & Jacquelin, 2 Cir., 391 F.2d 555 (1968), often referred to as Eisen II. On that appeal we remanded the case to the District Court for reconsideration and for findings on specific issues and we retained jurisdiction.1 While entertaining grave doubts on the questions of notice and manageability, we thought the original rejection of the case as a class action had been too summary, that improper standards had been applied and inadequate consideration given to the specific requirements of amended Rule 23.
After almost five years the case is before us again. While much of the delay *1008is not attributable to further proceedings bearing on issues arising under amended Rule 23, a very considerable amount of the time of Judge Tyler and of the lawyers for the respective parties, has, in this interval of five years, been devoted to hearings, the taking of depositions, and the preparation and filing of various District Court opinions, preceded by briefs and extensive oral arguments. Much of this time was devoted to an effort by Eisen’s counsel to meet the apparently insurmountable difficulties of notice and manageability by adopting the erroneous and frustrating view that some way must be found to make the ease viable as a class action. In the end Judge Tyler was persuaded that the innovations described by Judge Weinstein in his speeches2 and in his series of opinions in Dolgow v. Anderson, D.C., 43 F.R.D. 21 (1967); 43 F.R. D. 472 (1968); 45 F.R.D. 470 (1968); 53 F.R.D. 661 (1971); 53 F.R.D. 664 (1971), were authorized by amended Rule 23. These innovations were the preliminary mini-hearing on the merits and the “fluid recovery,” both of which will be fully described in due course. It is clear to us that, with or without these innovations, the notice provided by amended Rule 23 to be given “to all members (of the class) who can be identified through reasonable effort” cannot be given, as Eisen refuses to pay or put up any bond to cover this expense, and, if defendants prevail on the merits, they will be unable to recover any amounts expended by them for this purpose. We are also of the opinion that, on the basis of the new evidence now before us, the lawsuit is unmanageable as a class action, and that no preliminary mini-hearing on the merits and no “fluid recovery” procedures are authorized by the text or by any reasonable interpretation of amended Rule 23. Accordingly, we reverse and dismiss the case as a class action. We also vacate the findings of fact and conclusions of law that were made after the preliminary mini-hearing on the merits.
I
The Decision Below — 52 F.R.D. 253 1971
In 1968, when the case was previously before us, it was estimated by someone that there were 3,750,000 members of the class, consisting of those who had bought or sold odd lots on the New York Stock Exchange in the period from May 1, 1962 through June 20, 1966. It was then doubtful whether any of the members of the class could be “identified through reasonable effort.” Eisen’s position then was and now is that, except possibly in the eventuality of the ultimate adoption of Judge Tyler’s suggested plan which envisages payment by defendants of 90% of the cost of giving notice, he will not defray any of the expense of giving notice to any of the members of the class, nor will he post any bond to reimburse defendants for any of their disbursements, pursuant to any order of the District Court, made for the purpose of giving any notice.
It now appears that there are 6,000,000 members of the class and of these 2,250,000 can be easily identified.3 Members of the class reside in every state of the United States and most for*1009eign countries. They speak and understand a great variety of modern languages. The damages sought to be recovered were estimated at the time we last considered the case at something between a maximum of $60,000,000 and a minimum of $22,000,000.4 Now the estimate has been raised by Eisen’s counsel to 120 millions of dollars.
In our prior opinion we stated unequivocally that actual notice must be given to those whose identity could be ascertained with reasonable effort and that “in this type of case” plaintiff must pay the expense of giving notice to these members of the class.5 We further stated that if this could not be done there might be no other alternative than the dismissal of the case as a class action. For some reason not clear to us Judge Tyler disregarded these holdings and concluded that he had discretion, even with reference to those members of the class who could be easily identified, to provide for such notice as he thought to be reasonable in the light of the facts of this particular case.
Thus he directed actual notice only to “the approximately 2000 or more class members who had ten or more transactions during the relevant period” and to “5000 other class members selected at random” from the 2,500,000 class members who could easily be identified.6 With respect to the rest of the 6,000,000 members of the class, Judge Tyler ordered what, without reciting all the details concerning the schedule of proposed publications, we consider to be a totally inadequate compliance with the notice requirements of amended Rule 23. One of the reasons for this was perhaps because Judge Tyler thought of these first notices, by mail and by publication, as merely the first of a series of notices. Judge Tyler then deferred the question of who should pay for this first round of notices until after a “brief” preliminary hearing on the merits. This is what is called the “mini-hearing.” We *1010shall have more to say later about this preliminary mini-hearing on the merits of Eisen’s triple damage antitrust claim. Accordingly, the hearing was held “on the issue of the allocation of the costs of notice” and Judge Tyler concluded that the defendants must bear 90% of these expenses.
To describe Judge Tyler’s general scheme as it slowly developed in the series of his many opinions 7 following the remand would be too tedious. The sum and substance of it was that he at last realized that it was highly improbable that any great number of claims would, for a variety of reasons, ultimately be filed by the 6,000,000 members of the class. No claimant in the 6 years of the progress of the action had shown any interest in Eisen’s claim. The average odd-lot differential on each transaction had been $5.18. The average individual class member engaging in five transactions would have paid a total odd-lot differential of $25.90. Assuming a 5% illegal overcharge the recovery is approximately $1.30, and when trebled the average class member would be entitled to damages of $3.90. As the costs of administration might run into the millions of dollars, it was not likely that a rush of claimants would eventuate no matter how extensive the publication. As he had surmised in the beginning, and as Chief Judge Lumbard stated in his dissent (Eisen II, 391 F.2d p. 571), the class action was hopelessly unmanageable. So Judge Tyler tried to pull the case out of this morass by resorting to the “fluid recovery,” which had been used as a vehicle for carrying out a voluntary settlement in the Drug Cases, State of West Virginia v. Chas. Pfizer & Co., Inc., et al., 314 F.Supp. 710 (S.D.N.Y.1970).8
The concept of this “fluid recovery” is very simple. Having decided that there is no conceivable way in which any substantial number of individual claimants can ever be paid, “the class as a whole” is substituted for the 6,000,000 claimants. Thus the first round of notices becomes relatively unimportant. The scheme adopted envisages the first round of notices as sufficient to get the ball rolling. Little is said about Step Two. This involves a trial of the case to a judge and jury on the merits — not a preliminary mini-trial this time, but a real full scale trial of the private triple damage antitrust case. In some way the damages to “the class as a whole” will be assessed and the defendants, it seems to be assumed, will promptly pay this huge sum into court. This sum is supposed to constitute the “gross damages” to “the class as a whole.” With the money in hand, the case begins to resemble the Transitron and Drug Cases and from then on we are to have the real notices soliciting the filing of claims, the processing of these claims, the fixing of counsel fees and the payment of the general expenses of administration. As “the class as a whole” will include all those who had purchased or sold in the period from mid-1962 to mid-1966 and all those who, at the time of assessing the full damages, were presently purchasing or selling, and those who might in the future purchase and sell, securities in lots of less than 100 shares, it is quite apparent that some of the original 6,000,000 claimants will receive nothing, because they have never heard of the case or for other reasons have failed to file claims and have them processed, and many other new traders, who had no transactions in the period from mid-1962 to mid-1966, will receive some payments. According to Judge Tyler, at least those members of the original class of 6,000,000 who “have maintained their odd-lot activity, will reap the benefits of any recovery” (52 F.R.D. at p. 265). As far as we are aware there has never been, nor can there ever be, a reliable or even rational estimate of how many traders, whether *1011speculators or investors, can be said to be expected to continue as such after the lapse of 10 years or so. As it is suspected that relatively few claims will be filed and the damages assessed are supposed to cover the losses of “the class as a whole,” there will be a huge residue, similar to the amounts paid to various charities “to advance public health projects” in the Drug Cases, and this residue is to be used for the benefit of all. odd-lot traders by reducing the odd-lot differential “in an amount determined reasonable by the court until such time as the fund is depleted.” 52 F.R.D. at p. 265. We are at a loss to understand how this is to be done, but it is suggested that it “might properly be done under SEC supervision or at least with SEC approval.”
Despite its early expression of doubt on the subject,9 and what appears at least to be de facto exercise of supervisory powers over Rules or practices on the subject of the amount and uniformity of the rate of commissions on odd-lot purchases and sales, the SEC finally did exercise such powers10 and we hold that at all times the SEC possessed such powers under Sections 11(b)11 and 19(b)12 of the Securities Exchange Act of 1934. No District Court has authority to decide upon the rate of such commissions to take effect until the exhaustion of any residual fund left over by application of a “fluid recovery” in a private triple damage antitrust case, after the payment of claims in a class action or otherwise. The courts may review rulings of the SEC, but they have no more power than the District Court to fix any such rates in the first place or to give directions to the SEC concerning the fixing of such rates or the time within which such rates are to be effective, as part of a judgment in a private triple damage antitrust case.
II
Disposition of Certain Contentions of the Parties
Solely for the purpose of making our holdings clear in this difficult and complicated case we think it proper first to dispose of certain contentions of the parties.
A
We must reject Eisen’s claim that the fluid class recovery theory is not ripe for review. Indeed, there is no way to side-step this issue. We specifically remanded the case for consideration of the problem of manageability. The further proceedings on the remand were necessarily concerned with ascertaining whether there was a judicially sound way effectively to administer this action. Administration, of course, includes proof of damages and the distribution of the same. As we point out later in this opinion, Eisen concedes that the action is not manageable if fluid class recovery is not permissible. We *1012must face this issue if we are to pass on the question of manageability, which is the most important point in the case. We are no longer at the early stages of this case where it might be possible to put off to a later time the troublesome question of what to do with the damage fund if only a small number of claims are filed against the fund. See In Re Antibiotic Antitrust Actions, 333 F. Supp. 278, 281-282 (S.D.N.Y.1971).
B
Moreover, we think the three cases cited by Judge Tyler as “respectable precedent” for fluid class recovery are all distinguishable. These three cases are: Bebchick v. Public Utilities Commission, 115 U.S.App.D.C. 216, 318 F.2d 187, cert. denied, 373 U.S. 913, 83 S.Ct. 1304, 10 L.Ed.2d 414 (1963); the Drug Cases, 314 F.Supp. 710 (S.D.N.Y.), aff’d 440 F.2d 1079 (2d Cir. 1971); and Daar v. Yellow Cab Company, 67 Cal.2d 695, 63 Cal.Rptr. 724, 433 P.2d 732 (1967).
Judge Wyatt’s extraordinary feat of judicial administration in carrying out the terms of the one hundred million dollar settlement in the Drug Cases deserves all the praise it has received. But it was a consensual affair made possible by the agreement of the parties and without objection to the assumption by the District Court of jurisdiction to accept and administer the fund. Here we have no fund. There is no settlement. Every issue is contested and litigated. And authority to permit this action to proceed as a class action must be found within the four corners of amended Rule 23, as interpreted in the Reviser’s Note. Applying this test we hold Eisen’s class action must be dismissed. Bebchiek was not a class action in any sense of the word. Amended Rule 23 was not involved. In the exercise of its powers of review, the Court of Appeals for the District of Columbia Circuit reversed a judgment of the District Court approving the action of the Public Utilities Commission of the District of Columbia supporting a fare increase by the transit company. In the meantime, the additional cash fares, now found to be illegal, had been collected. There was no way to direct refunds as those who paid these cash fares could not be identified. So, also in the exercise of its powers of review the Court of Appeals directed the amount of these additional cash fares to be set up in the books of the transit company to be used, in the discretion of the regulatory commission “to benefit bus riders as a class in pending or future rate proceedings.” We cannot find that this case has any bearing on any of the issues in this amended Rule 23 case. Finally, Daar was a case arising under a state class action statute very different in its phraseology from amended Rule 23. The ruling was made on a demurrer to the complaint so the approach to the legal issues was entirely different from the making of a judicial determination, on the basis of proof, of whether or not the requirements of amended Rule 23 had been met. Moreover, the court was evidently of the view that the individuals who had been damaged by the alleged overcharge in taxi fares would ultimately have to prove their separate and individual damages. 433 P.2d at 740.
For the reasons stated in this opinion we disagree with the holding in the Dol-gow cases.
C
The Merits of Eisen’s Triple Damage Antitrust Claim
The defendants by extensive and even cogent arguments have urged us not only to vacate Judge Tyler’s findings and conclusions on the merits of the case but to decide that there is no merit in Eisen’s antitrust claim. Thus the defendants argue that in the context of an antitrust suit brought against defendants acting within a self-regulatory industry the per se rule of antitrust liability is inapplicable, and that even under a rule of reason the practices of the odd-lot defendants and the Exchange *1013were not violative of the antitrust laws. In essence the defendants contend that Judge Tyler incorrectly applied the Silver v. New York Stock Exchange, 373 U.S. 341, 83 S.Ct. 1246, 10 L.Ed.2d 389 (1963) doctrine, and should have concluded that the fixing of the odd-lot differential is necessary to the proper implementation of the Securities Act. Also the defendants assert that the Securities and Exchange Commission has primary jurisdiction over the substantive issues presented in this litigation. As already appears in this opinion, we hold and decide that the SEC does possess supervisory powers over the amount and uniformity of the commissions to be paid on odd-lot purchases and sales. But this is only part of the picture. There are many facets to this complicated case. We feel it is proper for us to say only that the record before us constitutes no proper basis for any decision of the merits, tentative or otherwise.
Ill
Controlling Principles
When discoursing on the arts or belles-lettres colorful language stimulates the imagination, beguiles one into useful symbolism and opens up the avenues to creative thought. But in the process of rationalizing legal conclusions and arriving at a sound and proper determination of questions of the interpretation of statutes, procedural rules and constitutional limitations, clichés and rhetorical devices generally miss the mark. Something more substantial is necessary to establish a base for the proper decision of difficult and complex questions of law. One reason for this is that the solution is found more often than not by the application of fundamentally simple principles.
Thus statements about “disgorging” sums of money for which a defendant may be liable, or the “prophylactic” effect of making the wrongdoer suffer the pains of retribution and generally about, providing a remedy for the ills of mankind, do little to solve specific legal problems. The result of this approach is almost always confusion of thought and irrational, emotional and unsound decisions. In cases involving claims of money damages all litigation presumes a desire on the part of the judicial establishment to make the wrongdoer pay for the wrongs he has committed, but to do this by applying settled or clearly stated principles of law, rather than by some process of divination. Punishment of wrongdoers is provided by law for criminal acts in statutes making it a crime punishable by fine or imprisonment to violate the antitrust laws. In certain civil suits punitive damages may be awarded; and in private antitrust cases the possible recovery of triple the loss actually suffered by a plaintiff is very properly praised as a supplementary deterrent. But none of these considerations justifies disregarding, nullifying or watering down any of the procedural safeguards established by the Constitution, or by congressional mandate, or by the Federal Rules of Civil Procedure, including amended Rule 23. It is a historical fact that procedural safeguards for the benefit of all litigants constitute some of the most important and salutary protections against oppressions,13 including oppressions by those whose intentions may be above reproach.
We adhere to what we have written in support of the remand of this *1014case in Eisen II. On the basis of the new evidence .adduced on the remand, what we are now doing is interpreting and applying various provisions of an amended and improved procedural device intended to facilitate the judicial disposition of the individual claims of the separate members of a class of persons so numerous that joinder of all members is impracticable. Amended Rule 23 was not intended to affect the substantive rights of the parties to any litigation. Nor could it do so as the Enabling Act that authorizes the Supreme Court to promulgate the Federal Rules of Civil Procedure provides that “such rules shall not abridge, enlarge or modify any substantive right.”14
The applicable substantive law is Section 4 of the Clayton Act15 that authorized the private triple-damage antitrust suit to recover damages by a person who has been “injured in his business or property” by reason of a violation of the antitrust laws. That the claims of many may not be treated collectively or as “the class as a whole” is what the Supreme Court decided in Snyder v. Harris, 394 U.S. 332, 89 S.Ct. 1053, 22 L.Ed.2d 319 (1969), where plaintiff, in a diversity case sued as representative of a class of some 4000 shareholders of an insurance company. Her individual claim was for less than $10,000. She was not permitted to aggregate her claim with the separate claims of the other members of the class which amounted to $1,200,000. Despite the fact that amended Rule 23 was aleady in effect, the case was dismissed. Moreover, in the recent case of Hawaii v. Standard Oil Co. of California, et al., 405 U.S. 251, 92 S.Ct. 885, 31 L.Ed.2d 184 (1972),16 it was again held that only persons actually injured in their business or property could claim damages under the Clayton Act.
Eisen also alleges that the Exchange is liable for its failure to regulate the odd-lot differential as required by Section 6 of the Securities Exchange Act. He argues that Section 6, 15 U.S.C., Section 78f, requires the securities exchanges to prescribe rules and regulations for the regulation of the industry. Failure so to regulate allegedly subjects the defendant-Exchange to liability to those who have suffered injury due to the abdication of this regulatory responsibility. The duty to regulate emanates from the Exchange Act, but the right of an injured party to recover damages is, according to Eisen, based upon federal common law, J. I. Case Co. v. Borak, 377 U.S. 426, 84 S.Ct. 1555, 12 L.Ed.2d 423 (1964); Baird v. Franklin, 141 F.2d 238 (2d Cir.), cert. denied, 323 U.S. 737, 65 S.Ct. 38, 89 L.Ed. 591 (1944). Thus, if Eisen is correct and Section 6 of the Act creates a statutory -duty on the Exchange to protect members of plaintiff’s class, then these members “may sue for injuries resulting from its breach and (the) common law will supply a remedy if the statute gives none,” Baird v. Franklin, 141 F.2d at 245.
The fundamental doctrine that permeates our opinion in Eisen II, and which we are now about to apply again in this same case, is whether the requirements of amended Rule 23 have been met. We find no helpful analogy in the procedures that have been used for generations in connection with preliminary injunctions or other provisional remedies intended to preserve the status quo.
*1015So, we shall proceed to examine in some detail the provisions of amended Rule 23 in the light of the long and explicit Advisory Committee’s Note, 39 F. R.D. 98, and decide whether or not Judge Tyler has followed our directions to appraise and to apply each of the factors enumerated on the face of the Rule. In the light of our conclusions with respect to notice and manageability, we do not reach the subject of adequate representation.
IV
Lack of Individual Notice to “All Members Who Can be Identified Through Reasonable Effort”
Our prior ruling in Eisen II is clear and specific. If identification of any number of members of the class can readily be made, individual notice to these members must be given and Eisen must pay the cost. If this cannot be done, the case must be dismissed as a class action. Amended Rule 23(c)(2) unambiguously states that notice to the class generally shall be the “best notice practicable,” and then “including individual notice to all members who can be identified through reasonable effort.” Moreover, the Advisory Committee’s Note states (39 F.R.D. 106-107): “Indeed, under subdivision (c)(2), notice must be ordered, it is not merely discretionary * * 17 While Judge Tyler seems to have realized that this phase of amended Rule 23 has decided constitutional overtones, he apparently thought the flexibility of the Rule and our statement that the Rule was to be given a liberal interpretation authorized him to exercise his discretion even if this involved the complete disregard of our specific and unambiguous ruling on the subject of actual individual notice to identifiable members of the class. This ruling alone compels a reversal of the order appealed from and the dismissal of the case as a class action.
V
The Preliminary Mini-Hearing on the Merits Weis' Not Authorized by Amended Rule 23 and the District Court Had No Jurisdiction or Competence to Hold Such a Hearing
The Federal Rules of Civil Procedure set forth a considerable variety of procedural devices designed for the disposition of cases on the merits. There may be traditional trials to a judge or to a judge and jury; there may be summary judgments, dismissals with or without prejudice for failure to state a claim and so on. But neither in amended Rule 23 nor in any other rule do we find provision for any tentative, provisional or other makeshift determination of the issues of any ease on the merits for the avowed purpose of deciding a collateral matter such as which party is to be required to pay for mailing, publishing or otherwise giving any notice required by law. In most cases the so-called tentative findings and conclusions arrived at without the salutary safeguards applicable to all full scale trials on the merits will be extremely prejudicial to one or the other of the parties who bear the brunt of such findings and conclusions, and such prejudice may well be irreparable.
*1016We agree with the ruling by the Fifth Circuit in Miller v. Mackey International, Inc., 452 F.2d 424 (5th Cir. 1971), that the preliminary hearing on the merits was improper. As stated by Judge Wisdom, 452 F.2d at page 427:
In determining the propriety of a class action, the question is not whether the plaintiff or plaintiffs have stated a cause of action or will prevail on the merits, but rather whether the requirements of Rule 23 are met.
This was the first time, as far as we are aware, that any Court of Appeals has passed on the point. As noted by Judge Wisdom in footnote 5 on page 429 of 452 F.2d, this Court did not have occasion to rule on the question when it decided the appeal from Judge Weinstein’s summary judgment for defendants in Dolgow, 438 F.2d 825 (2d Cir., 1971). While Judge Weinstein’s oral order not only granted summary judgment for defendants but also held the case not to be a proper class action, the posture of the case on the appeal to our Court was such that the Court had no occasion to consider the propriety of the preliminary mini-hearing on the merits that had been conducted by Judge Weinstein. Nor was that question before this Court in Green v. Wolf Corporation, 406 F.2d 291 (1968). See footnote 15 on pages 301-302.
Of the few District Court decisions on the point most of these disagree, as, of course, does the Fifth Circuit, with the innovations described in Dolgow,18 and there is little to commend the reasoning or lack of reasoning in the others.19 No provision is made in amended Rule 23 for any such mini, preliminary or other hearing on the merits. It does violence to the whole «concept of summary judgment, and cannot be reconciled with the requirement in Rule 23 that “as soon as practicable after the commencement of the action” the question of "class suit vel non be decided.
Moreover, in this case we did not in our disposition of the prior appeal intend to relinquish to the District Court any jurisdiction to pass on the merits of the case but only to decide if the requirements of amended Rule 23 had been met. Accordingly, we are constrained to hold that the whole preliminary mini-hearing on the merits proceeding, including the findings of fact and conclusions of law, was conducted and made without jurisdiction.
VI
As a Class Action the Case Is Unmanageable
From the beginning it has been Judge (then Chief Judge) Lumbard's view that as a class action the ease is unmanageable and that it should be dismissed as a class action. It turns out that he was right. As soon as the evidence on the remand disclosed the true extent of the membership of the class and the fact that Eisen would not pay for individual notice to the members of the class who could be identified, and the evidence fur*1017ther disclosed that the class membership was of such diversity and was so dispersed that no notice by publication could be devised by the ingenuity of man that could reasonably be expected to notify more than a relatively small proportion of the class, a ruling should have been made forthwith dismissing the case as a class action. This dismissal could have saved several years of hard work by the judge .and the lawyers and wholly unnecessary expense running into large figures. The fact that the cost of obtaining proofs of claim by individual members of the class and processing such claims was such as to make it clear that the amounts payable to individual claimants would be so low as to be negligible also should have been enough of itself to warrant dismissal as a class action. Other cases involving millions of diverse and unidentifiable members of an alleged class had been dismissed as unmanageable or altered in composition.20 And so even Eisen and his counsel conceded that the class was not manageable unless the “fluid recovery” procedures were adopted.
Thus, in the language of Eisen’s counsel:
There are some six million persons in the class. If each had to present his own personal claim for damages, the class, indeed, would not be manageable. The facts in Stipulation No. 2 only underscore the obvious. In both Cherner v. Transitron Electronic Corporation, 201 F.Supp. 934 (D. Mass.1962) and Illinois Bell Telephone Co. v. Slattery, 102 F.2d 58 (7th Cir. 1939). (the eases included in Stipulation No. 2), the refund process overwhelmed the refunds. And both cases involved, quite obviously, classes much smaller than the class at bar.
Where there are millions of dispersed and unidentifiable members of the class notices by publication giving the essential information required by amended Rule 23 are a farce.21 And, when it comes to the filing and processing of claims, lawyers specializing in class actions have stated that the only effective way to induce any reasonable number of members of the class to file claims is to conduct full-scale campaigns on TV and radio, solicit appearances by advocates of consumers’ rights such as Ralph Nader, letters from Congressmen to their constituents,. public statements by various state attorneys general “and coverage in various news media, union newsletters and the like,” also to persuade the Federal Communications Commission to classify announcements of this character as “public service announcements.” 22
All the difficulties of management are supposed to disappear once the “fluid recovery” procedure is adopted. The claims of the individual members of the class become of little consequence. If the damages to be paid were only the aggregate of the sums found due to individual members of the class, after their claims had been processed, it is *1018fairly obvious that in cases like Eisen the expenses of giving the notices required by amended Rule 23 and the general costs of administration of the action would exceed the amount due to the few members of the class who filed claims and the individual members of the class would get nothing.23 We referred to this possibility in our opinion in Eisen II, 391 F.2d at pages 567 and 570.
But if the “class as a whole” is or can be substituted for the individual members of the class as claimants, then the number of claims filed is of no consequence and the amount found to be due will be enormous, affording, we are told, plenty of money to pay all expenses, including counsel fees, and a residue so large as to justify reduction of the odd-lot differential for years in the future, for the benefit of all traders, past, present and future, who are to be considered to be members of “the class as a whole.”
Even if amended Rule 23 could be read so as to permit any such fantastic procedure, the courts would have to reject it as an unconstitutional violation of the requirement of due process of law. But as it now reads amended Rule 23 contemplates and provides for no such procedure. Nor can amended Rule 23 be construed or interpreted in such fashion as to permit such procedure. We hold the “fluid recovery” concept and practice to be illegal, inadmissible as a solution of the manageability problems of class actions and wholly improper.
VII
A Few General Observations
Perhaps in part due to the liberal views on the subject of amended Rule 23, as expressed in our opinion in Eisen II, and doubtless stimulated by the counsel fees allowed in the Transitron and the Drug Cases, where large voluntary settlements had been approved and administered by District Judges, there has followed such a quantity of comment pro and con on the questions of law we are to decide in this case, by law professors, by judges, and especially by lawyers specializing in class actions, expressed in numerous articles, opinions and published speeches, that the task of even attempting to enumerate all of these for purposes of documentation is too much for us.
Class actions have sprouted and multiplied like the leaves of the green bay tree. No matter how numerous or diverse the so-called class may be or how impossible it may be ever to compensate the individual members of the class, a champion steps forth. Thus class actions have been brought “on behalf of all subscribers of business telephones in New York County, all Master Charge credit card holders similarly situated, all consumers of gasoline in a given state or states, all homeowners in the United States, and even all people in the United States.” 24 So far as we are aware not a single one of these class actions including millions of indiscriminate and unidentifiable members has ever been brought to trial and decided on the mer*1019its. But the preliminary procedures, including the preliminary mini-hearing on the merits, such as those conducted by Judge Tyler in order to decide whether or not this case was a proper class action, and the huge and unavoidable expense of producing witnesses and documents pursuant to discovery orders, have brought such pressure on defendants as to induce settlements in large amounts as the alternative to complete ruin and disaster, irrespective of the merits of the claim.25
The “in terrorem” effects of the innovations described in Dolgow have been highly praised by those who invented or applied them.26 But Professor Milton Handler, whose Annual Antitrust Review has for many years brought his expertise in Trade Regulation and, we are happy to say, some entertainment to the members of the Association of the Bar of the City of New York, and to the members of the Bench and Bar in general, minces no words. He calls these procedures “legalized blackmail.”27 There is reason to believe that the practical effect of these procedures, and the fact that possible recoveries run into astronomical amounts, generate more leverage and pressure on defendants to settle, even for millions of dollars, and in cases where the merits of the class representatives claim is to say the least doubtful, than did the old-fashioned strike suits made famous a generation or two ago by Clarence H. Venner.
And yet, even if amended Rule 23 furnishes no satisfactory solution in situations where immense numbers of consumers have been mulcted in various ways by illegal charges, it would seem that some means should be provided by law for the redress of these wrongs to the community and to society as a whole. The numerous decisions by courts in these class action cases have at least exposed the lack of adequate remedy under existing laws. From our extensive study of the whole situation in working on this Eisen case it would seem that amended Rule 23 provides an excellent and workable procedure in cases where the number of members of the class is not too large. It seems doubtful that further amendments to Rule 23 can be expected to be effective where there are millions of members of the class, without some infringement of constitutional requirements. The problem is really one for solution by the Congress. Numerous administrative agencies protect consumers in various ways. It should, we think, be possible for the Congress to create some public body to do justice in the matter of consumers’ claims in such fashion as to afford compensation to the injured consumer. If penalties are to be imposed upon wrongdoers, at least let the Congress decide how the money is to be spent.
Another possibility, suggested by the Report and Recommendations of the Special Committee of the American College of Trial Lawyers, is a further amendment to amended Rule 23 consisting of a new subdivision providing:
In an action commenced pursu'ant to subdivision (b)(3), the court shall consider whether justice in the action would be more effectively served by maintenance of the action as a class *1020action pursuant to subdivision (b)(2) in lieu of (b) (3).
The procedure involved in applying for prospective injunctive relief is relatively simple and inexpensive, social and economic reforms may be implemented and an end put to illegal practices with far more benefit to the community than that derived from minimal or token payments to individual members of a class. Attorney’s fees in such c^ses should also provide adequate incentive to counsel for the representative or representatives of the class.28
Conclusion
For the reasons stated in this opinion the findings and conclusions following the mini-hearing are vacated and set aside, the various rulings of the District Court sustaining the prosecution of the case as a class action are reversed and, as a class action, the case is dismissed, without prejudice to the continuance of so much of the claim asserted in the complaint as refers to Eisen’s alleged individual rights against the defendants.
HAYS, Circuit Judge (concurring in the result):
I concur in the result because I am unable to accept the ruling of the district court requiring the defendants to pay 90 per cent of the cost of notice, since, if the defendants should finally prevail, they would not be reimbursed for this expenditure.
ON PETITION FOR REHEARING.
A petition for a rehearing having been filed herein by counsel for the appellee,
Upon consideration thereof, it is
Ordered that said petition be and it hereby is denied.
A petition for a rehearing containing a suggestion that the action be reheard en banc having been filed herein by counsel for plaintiff-appellee, a poll of the judges in regular active service having been taken at the request of such a judge, and there being no majority in favor thereof,
Upon consideration thereof, it is
Ordered that said petition be and it hereby is denied.
Judges Hays, Oakes and Timbers dissent.
/s/ HENRY J. FRIENDLY Chief Judge
KAUFMAN, Circuit Judge, with whom FRIENDLY, Chief Judge, and FEIN-BERG, MANSFIELD, and MULLIGAN, Circuit Judges, concur.
I vote against en banc, not because I believe this case is unimportant, but because the case is of such extraordinary consequence that I am confident the Supreme Court will take this matter under its certiorari jurisdiction. Judge Oakes’s opinion, dissenting from the denial of en banc, illustrates some of the far-reaching implications the panel’s opinion might have on the initiation and administration *1021of certain class action litigation in the future. En banc consideration by this court, however, would merely serve as an instrument of delay. Moreover, the application for certiorari will not go to the Supreme Court barren of the views of the judges of this court as, for example, in the Pentagon Papers case, where the court convened en banc but, because of urgent time considerations, did not write opinions. Judge Oakes has set forth his views on the merits with vigor and Judge Medina’s panel opinion articulates the opposing position. Our decision to decline en banc consideration of this case in no way implies, as my brother Oakes suggests, the demise of en banc in future cases of exceptional importance; nor does it threaten to turn this collegial court into a fragmented judicial body of panels of three, in which each panel’s opinions speak only for the panel, and not for the whole Court. Instead, we wisely speed this case on its way to the Supreme Court as an exercise of sound, prudent and resourceful judicial administration.
*1020Something seems to have gone radically wrong with a well-intentioned effort. Of course, an injured plaintiff should be compensated, but the federal judicial system is not adapted to affording compensation to classes of hundreds of people with $10 or even $50 claims. The important thing is to stop the evil conduct. For this an injunction is the appropriate remedy, and an attorney who obtains one should be properly compensated by the defendant, although not in the astronomical terms fixed when there is a multimillion dollar settlement. If it be said that this still leaves the defendant with the fruits of past wrong-doing, consideration might be given to civil fines, payable to the government, sufficiently substantial to discourage engaging in such conduct but not so colossal as to produce recoveries that would ruin innocent stockholders or, what is more likely, produce blackmail settlements. This is a matter that needs urgent attention.
. We retained jurisdiction because of the doubt engendered by the phraseology of the “death knell” opinion, 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) (Eisen I). We thought this case might be construed as making appealable an interlocutory order dismissing a ease as a class action because “no lawyer of competence is going to undertake this complex and costly case to recover $70 for Mr. Eisen,” see Korn v. Franchard Corp., 443 F.2d 1301 (2d Cir. 1971); Green v. Wolf Corp., 406 F.2d 291 (2d Cir. 1968), cert. denied, 395 U.S. 977, 89 S.Ct. 2131, 23 L.Ed.2d 766 (1969), but perhaps denying appealability to a similar order holding the case to be a proper class action under amended Rule 23. While the “death knell” doctrine has been criticized by the Third Circuit in Hackett v. General Host Corp., 455 F.2d 618 (3d Cir.), cert. denied, 407 U.S. 925, 92 S.Ct. 2460, 32 L.Ed .2d 812 (1972), we think the consequences of class action rulings are so serious that these interlocutory orders should be made appealable, provided we adopt the view expressed by Chief Judge Friendly in Korn v. Franchard Corp., supra, 443 F.2d at 1307, to the effect that this Court should formulate the rule of appealability in such fashion that it “will afford equality of treatment as between plaintiffs and defendants.” The same considerations which led this Circuit to apply the rule of Cohen v. Beneficial Industrial Loan Corp., 337 U.S. 541, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949) in Eisen I, also would seem to require a rule allowing a defendant to appeal from an interlocutory order permitting the representative plaintiff to continue the suit as a class action.
The “collateral order” doctrine of Colien is based on the pragmatic view that a decision which finally determines an issue in the case which is crucial to the further conduct of the case, and is collateral to the merits of the action, is to receive immediate appellate review if delay in such review will cause “irreparable harm” to the complaining party. The seeds of Cohen were sowed by the Supreme Court as early as Forgay v. Conrad, 6 How. 201, 205, 12 L.Ed. 404 when it was said that appealability should be allowed if the effect of an interlocutory order is such that if the order is immediately carried into execution the defendant “may be ruined before he is permitted to avail himself of the right” to appeal. An order sustaining a class action allegation clearly involves issues “fundamental to the further conduct of the case;” (see Gillespie v. United States Steel Corp., 379 U.S. 148, 85 S.Ct. 308, 13 L.Ed.2d 199 (1964); Larson v. Domestic and Foreign Commerce Corp., 337 U.S. 682, 69 S.Ct. 1457, 93 L.Ed. 1628 (1949) ; United States v. General Motors Corp., 323 U.S. 373, 65 S.Ct. 357, 89 L.Ed. 311 (1945)) ; the order is also separable from the merits of the caso; and irreparable harm to a defendant in terms of time and money spent in defending a huge class action when an appellate court may years later decide such an action does not conform to the requirements of Rule 23, is evident. By this extension or interpretation of Eisen I the rule of finality would be given a “practical rather than a technical construction,” see Cohen, supra, 337 U.S. at 546, 69 S.Ct. at 1226; Gillespie v. United States Steel Corp., supra, 379 U.S. at 152, 85 S.Ct. at 311; and we would avoid a possible denial of justice caused by delaying review by permitting an interlocutory appeal of rulings either sustaining or striking class action allegations. (See Swift & Company v. Compania Colombiana, 339 U. S. 684, 70 S.Ct. 861, 94 L.Ed. 1206 (1950); Dickinson v. Petroleum Corp., 338 U.S. 507, 70 S.Ct. 322, 94 L.Ed. 299 (1950)).
. Judge Weinstein’s speech on this subject was reported in Weinstein, “The Class Action Is Not Abusive,” New York Law Journal, May 1, 1972, p. 1, and May 2, 1972, p. 1.
. According to the opinion below, the names and addresses of approximately 2,000,000 class members can be identified. In addition, another 100,000 or more had odd-lot transactions in stocks listed on the Exchange through what is called the “Monthly Investment Plan.” These individuals can be identified through computer tapes in a manner similar to that used for locating the 2,000,000. Furthermore, another 150,000 or so public individuals had odd-lot transactions in stocks listed on the Exchange through “payroll deduction plans” operated by Merrill Lynch, Pierce, Fenner & Smith, Inc. Their names and addresses can be identified through the records of Merrill Lynch.
. 52 F.R.D. 253, 265 (S.D.N.Y.1971).
. We did not in our opinion in Eisen II intend our statement that “in such a case as this” plaintiff should defray the cost of notice to be dictum, as has been suggested, see Berland v. Mack, 48 F.R.D. 121, 131 (S.D.N.Y.1969). This statement was part of our interpretation of amended Rule 23 as applied in this case, especially with respect to the actual individual notice required by subdivision 23 (c) (2) to be given to those members of the class who could be identified. This was part of our instructions to conduct a hearing on the remand and decide whether the requirements of amended Rule 23 had been or could be met.
Nor did we decide or intend to say that in all eases or under all circumstances plaintiffs in class actions are or must be required to defray the cost of giving the various notices specified in amended Rule 23. This is an action to recover money damages for alleged violations of Section 4 of the Clayton Act and Section 6 of the Securities and Exchange Act of 1934. It is not a derivative stockholder’s action asserting a cause of action in favor of a defendant corporation, which regularly sends communications to all the stockholders and may be said to owe its stockholders certain fiduciary duties, nor a ease where a public utility corporation which regularly sends monthly bills to its current customers has been held to have overcharged its customers and the class suit is brought to compel a refund. There may be other similar examples of class actions in which, depending on the circumstances of particular cases, courts might find justification for holding that a representative plaintiff was not obligated to defray the cost of giving the notices required by amended Rule 23. We do not attempt any enumeration. It must be recalled that the provisions for notice in amended Rule 23 were intended to comply with constitutional requirements. See Advisory Committee’s Note, 39 F.R.D. 69, 107.
. Judge Tyler in discussing the notice problems observed that the plaintiff had also offered to send individual notice to all member firms of the New York Stock Exchange and to all commercial banks with large trust departments. This together with the individual notice to the 2000 class members with ten or more transactions and the 5000 selected at random, and the notice by publication, would in Judge Tyler’s view “increase the likelihood of reaching a significant portion of the class,” and would be in conformity with the requirements of due process.
. Judge Tyler’s opinions following the remand are reported at 50 F.R.D. 471, 52 F.R.D. 253, and 54 F.R.D. 565.
. Judge Wyatt’s approval of the settlement was affirmed by this Court at 440 F.2d 1079 (1971).
. According to the Report of the Special Study of Securities Markets, Vol. I, p. 182 (1963), an attorney representing the odd-lot firms presented a proposal for an increased differential. The Director of the Division of Trading and Exchange expressed some doubt as to the Securities and Exchange Commission’s jurisdiction over the matter. The Report states that “(t)he Commission declined to decide the jurisdictional question but stated that it had no objection to the proposal. It is to be noted that neither the Exchange nor the Commission asserted jurisdiction, the former denying its jurisdiction and the latter expressing doubts, yet each informally acquiesced in the increase.” See also p. 200 of the Report.
. At ’t|[ie preliminary hearing defendants’ Exhibit O contained a letter from the SEC to the New York Stock Exchange dated June 16, 1966. In part the letter stated that the “Commission hereby makes written request pursuant to section 19(b) of the Securities Exchange Act that the Exchange effect on its own behalf changes in its rules and practices in respect of odd-lot purchases and sales, and the fixing of reasonable rates of commission and other changes in connection therewith, to fix odd-lot differentials * *
. 15 U.S.C., Section 78k(b).
. 15 U.S.C., Section 78s (b).
. The procedural safeguards we speak of were wisely embodied in the Fifth Amendment’s Due Process Clause. The importance of due process of law and procedural fairness has been emphasized by some of our leading jurists. Justice Brandéis observed that “in the development of our liberty insistence upon procedural regularity has been a large factor.” (Burdeau v. McDowell, 256 U.S. 465, 477, 41 S.Ct. 572, 576, 65 L.Ed. 1048 (1921) (dissenting opinion)). Justice Frankfurter also remarked that “(f) airness of procedure is ‘due process in the primary sense * * *.’ It is ingrained in our national traditions.” (Joint Anti-Fascist Refugee Committee v. McGrath, 341 U.S. 123, 161, 71 S.Ct. 624, 643, 95 L.Ed. 817 (1951) (concurring opinion)).
. 28 U.S.C., Section 2072.
. 15 U.S.C., Section 15.
. The District Court in Hawaii v. Standard Oil Co., 301 F.Supp. 982 (D. Hawaii, 1969) dismissed the class action allegation on the ground that “under the circumstances * * *, the class action based upon the injury to every individual purchaser of gasoline in the State, * * in the context of the pleadings, would be unmanageable.” Hawaii, however, decided not to appeal this ruling to the Ninth Circuit, and, of course, the ruling was not before the Supreme Court for review in Hawaii v. Standard Oil Co., 405 U.S. 251, 256, n. 6, 92 S.Ct. 885, 31 L.Ed.2d 184 (1972).
. The Drug Cases, supra, 314 F.Supp. 710 (S.D.N.Y.1970), aff’d 440 F.2d 1079 (2d Cir. 1971), contain nothing that gives support to Judge Tyler’s ruling that individual notice to a limited number of members of a group of 2,250,000 whose names and addresses were identified was sufficient compliance with the notice requirements of amended Rule 23. An examination of the briefs in the Drug Cases makes it clear that reasonable effort would not have uncovered the names and addresses of the members of the consumer class of persons who bought the drugs on prescription at drug stores. Therefore, publication as to the consumer class was deemed sufficient. Moreover, as appellant in the Drug Cases attacked the sufficiency of the notice to the members of the wholesaler-retailer class, this Court rejected this argument, observing that individual notice was sent by mail to each and every member of this class whose name was available. 440 F.2d at 1091.
. Decisions which have rejected the use bf a preliminary hearing on the merits to decide the propriety of litigation proceeding as a class action include: Kahan v. Rosenstiel, 424 F.2d 161 (3d Cir.), cert. denied Glen Alden Corp. v. Kahan, 398 U.S. 950, 90 S.Ct. 1870, 26 L.Ed.2d 290 (1970); Katz v. Carte Blanche Corp., 52 F.R.D. 510 (W.D.Pa.1971); Fogel v. Wolfgang, 47 F.R.D. 213 (S.D.N.Y.1969); Cannon v. Texas Gulf Sulphur Co., 47 F.R.D. 60 (S.D.N.Y.1969); Mersay v. First Republic Corp. of America, 43 F.R.D. 465 (S.D.N.Y.1968). Judge Mansfield took the opportunity in Berland v. Mack, 48 F.R.D. 121, 132 (S.D.N.Y.1969) to comment on the prelim inary hearing on the merits: “The suggestion that such abuse of the corporate treasury can be avoided by a preliminary hearing to determine the merits of the claim is illusory. Quite aside from the additional burden that it heaps upon the Court, it would be a rare case where the Court could assure the class of ultimate success even after a preliminary hearing.”
. Cases supporting a preliminary hearing on the merits are Milberg v. Western Pacific Railroad Co., 51 F.R.D. 280 (S.D.N.Y.1970), appeal dismissed, 443 F.2d 1301 (2d Cir. 1971), and, of course, Dolgow v. Anderson, 43 F.R.D. 472 (E.D.N.Y.1968).
. See, City of Philadelphia v. American Oil Co., 53 F.R.D. 45 (D.N.J.1971); United Egg Producers v. Bauer International Corp., 312 F.Supp. 319 (S.D.N.Y.1970); Hackett v. General Host Corp., Civil No. 70-364 (E.D.Pa.1970), appeal dismissed, 455 F.2d 618 (3d Cir.), cert. denied, 407 U.S. 925, 92 S.Ct. 2460, 32 L.Ed.2d 812 (1972); Philadelphia Electric Co. v. Anaconda American Brass Co., 43 F.R.D. 452, 461 (E.D.Pa.1968).
. Notice by publication has been sanctioned as consistent with due process of law under certain circumstances, Mullane v. Central Hanover Bank & Trust Co., 339 U.S. 306, 70 S.Ct. 652, 94 L.Ed. 865 (1950). But in Schroeder v. City of New York, 371 U.S. 208, 212-213, 83 S.Ct. 279, 282, 9 L.Ed.2d 255 (1962) the Court refined the Mullane rule: “The general rule that emerges from the Mul-lane case is that notice by publication is not enough with respect to a person whose name and address are known or very easily ascertainable and whose legally protected interests are directly affected by the proceedings in question.”
. These suggestions were made and discussed in Shapiro, “Consumer Participation in Antitrust Class Action Part II”, New York Law Journal, May 31, 1972, p. 1.
. The recent draft for the Manual for Complex Litigation suggests that problems of administering the class action should not justify denial of an appropriate class action request “except when the attention and resources required to be devoted strictly to administrative matters will frustrate the securing of ultimate relief to which the class members may be entitled.” (p. 36). Due to our rejection of the fluid class recovery concept, we are compelled to conclude that this litigation falls squarely within this exception contemplated in the Manual. The administrative problems posed by this action will frustrate any effort to provide the individual class members with compensation for the alleged injuries.
. In the Report and Recommendations of the Special Committee of the American College of Trial Lawyers on Rule 23 of the Federal Rules of Civil Procedure, issued March 15, 1972, this quotation appears on p. 6 with supporting citations.
. Id., at 16. See also, Morris v. Burchard, 51 F.R.D. 530, 536 (S.D.N.Y.1971).
. See, Dolgow v. Anderson, 43 F.R.D. 472, 487 (E.D.N.Y.1968); Miller, Problem in Administering Relief In Class Actions Under Federal Rule 23(b) (3), 54 F.R.D. 501, 508; Pomerantz, New Developments in Class Actions—Has Their Death Knell Been Sounded?, 25 Bus. Lawyer 1259 (1970).
. Handler, The Shift From Substantive to Procedural Innovations in Antitrust Suits -The 23rd Annual Antitrust Review, 71 Colum.L.Rev. 1, 9 (1971). See also, Professor Handler’s 24th Annual Antitrust Review, 72 Colum.L.Rev. 1, 34—42, in which he criticizes the fluid recovery procedure as sought to be applied in “class” actions. There are some additional trenchant comments in his 25th Annual Antitrust Review, published in the December, 1972 issue of The Record of the Association of the Bar of the City of New York, pp. 660 ff.
. In his recent book Federal Jurisdiction : A General View, containing his 1972 Columbia University James S. Car-pentier Lectures, Chief Judge Friendly makes this comment on class actions pursuant to amended Rule 23, at page 120, omitting footnotes: