(dissenting).
I must respectfully dissent from the action taken by the majority.
Ours is an appellate, not a trial, court. And an appellate court has no authority to grant or to refuse a preliminary injunction. “It is to the discretion of the trial court and not to the appellate court, that the law has intrusted the power * * * to grant or dissolve an injunction, and the only question for an appellate court is: Does the proof clearly establish an abuse of that discretion by the trial court * * * for unless such abuse is clearly established, or an obvious error has occurred in the application of the law, or a serious and important mistake has been made in the consideration of the proof, the judgment of the trial court must be taken as presumptively correct.” Stokes v. Williams, 226 F. 148 (3 Cir.1915), cert. denied 241 U.S. 681, 36 S.Ct. 728, 60 L.Ed. 1234.
I am compelled to express this delineation of trial and appellate court functions because I fear that, in their approach to this problem my distinguished brothers have assumed the broad prerogatives of a trial court instead of confining themselves to the limited scope of review vested in the appellate structure. For an appellate court to reverse a trial court’s order refusing a preliminary injunction requires a determination that inappropriate law was applied to the facts or that clear error was committed in the finding of those facts. Neither circumstance becomes visible to me upon an examination of the trial record.1
*528One question of law will turn this case: whether there was a showing by Allis-Chalmers of a reasonable probability of success on final hearing. The trial judge found that plaintiff had not met its burden; the majority would reverse, and say that the burden was met.
Where pure questions of law are concerned, the scope of review of an appellate court is unlimited. But this is so only where the question of law “has an existence independent” of findings of fact. United States v. Yellow Cab Co., 338 U.S. 338, 70 S.Ct. 177, 94 L.Ed. 150 (1949). Where the ultimate question of law depends upon findings of fact, the scope of review becomes severely circumscribed ; an appellate court does not have carte blanche authority to pick and choose facts it would have found had it been the fact-finding tribunal; quite the contrary, it must accept the findings of fact of the trial judge unless there was clear error in the fact finding process.
These basic principles must govern our examination of the record and the action of the court below.
The majority’s action in reversing the trial court and ordering the issuance of a preliminary injunction is bottomed on a finding that the acquisition of Allis-Chal-mers by White Consolidated will in all probability enhance the power of a White-Allis combination to engage in illegal reciprocal dealing. In addition, my brother Stahl finds that a White-Allis combination will probably:
(a) eliminate potential competition in the metal rolling mill industry and other relevant markets, or
(b) diminish potential independent competition in other diversified markets.
None of these conclusions are based on pure questions of law with “an existence independent of findings of fact.” Instead, they are findings having characteristics of both law and fact — the ultimate conclusions resting on preliminary findings of fact. As such, law and fact are inextricably interwoven. Had the equity side of the court not been chosen by Congress to process private anti-trust litigation, the basic findings found by the majority to be significant enough to overturn the decision of the trial court would have been questions for a jury. That equity requires the fact finding to be made by a judge, and not a jury, does not dilute the exclusive factual characteristics of these findings.
This threshold consideration assumes critical proportions because the findings of fact by the trial judge must be adopted by the reviewing court unless set aside under the “clearly erroneous” rule. Fed.R.C.P. 52(a).
In the anti-trust case of United States v. Yellow Cab Co., supra, 338 U.S. at 342, 70 S.Ct., at 179, the Supreme Court discussed this very issue and said: “It ought to be unnecessary to say that Rule 52 applies to appeals by the Government as well as to those by other litigants. There is no exception which permits it, even in an antitrust case, to come to this Court for what virtually amounts to a trial de novo on the record of such findings as intent, motive and design. While, of course, it would be our duty to correct clear error, even in findings of fact, the Government has failed to establish any greater grievance here than it might have in any case where the evidence would support a conclusion either way but where the trial court has decided it to weigh more heavily for the defendants. Such a choice between two permissible views of the weight of the evidence is not ‘clearly erroneous’.”
*529In United States v. National Association of Real Estate Boards, 339 U.S. 485, 495-496, 70 S.Ct. 711, 94 L.Ed. 1007 (1950), Mr. Justice Douglas stated: “It is not enough that we might give the facts another construction, resolve the ambiguities differently, and find a more sinister east to actions which the District Court apparently deemed innocent, [citations omitted] We are not given those choices, because our mandate is not to set aside findings of fact ‘unless clearly erroneous’.”2
There is yet another reason why I am constrained to emphasize our role as an appellate court: in exercising our function of review, we are restricted to an examination of only those matters presented to the trial court. Our assessment must be limited to the four corners of the record which was before the lower court. As its name suggests, an appellate review confers neither jurisdiction nor prerogative to reopen a case on the merits or to consider matters dehors the record.
Reassertion of this basic postulate becomes necessary because a significant portion of the facts and legal conclusions advanced by the appellant are derived from the allegations contained in a proposed Federal Trade Commission complaint3 which was published after 'the *530decision of the lower court. This complaint is an ex-parte document disclosed by the FTC as the preliminary stage of possible governmental anti-trust action against White’s acquisition of Allis-Chal-mers.
I refuse to accept the quaint notion that the issuance of this “reason to believe” statement by the FTC has any legitimate place in the present appeal. My brother Stahl, nevertheless, while professing not to afford “excessive weight to the complaint,” has freely, and I believe, improperly, incorporated the facts and conclusions of this proposed complaint into the very foundations of his opinion. For example, over 30 per cent of the footnotes supporting the opinion refer to matters contained in the FTC complaint not to be found in the record below.4 I believe that such reliance on this extraneous material is unwarranted and unwise. There is no Congressional directive commanding it; there are almost two centuries of American judicial experience militating against it. For an appellate court to go beyond the record of a trial court and accept instead accusatory information of an agency of the executive branch of the government from an ex-parte preliminary proceeding is nothing .short of an abdication of traditional concepts of independent judicial review.
This is not to say that a district court in determining the issue of whether a reasonable probability of success has been shown, should, ostrich-like, totally ignore the pending of an FTC proceeding, even if it has not become “final” within the meaning of 15 U.S.C.A. § 26(a).5 Even then, however, the district court should not accept as gospel, either the conclusa-tory articulations of the prosecuting branch of this agency or facts which have not been tested in the crucible of an adversary proceeding.6
My quarrel, however, is not limited to this inordinate reliance upon matters not found in the trial record. Indeed, one of my major concerns stems from the cavalier treatment afforded those findings of fact made by the trial court. Not only were these findings not clearly erroneous, but in my view they were completely justified on the basis of the evidence presented.
The district court was confronted with gauging the anti-competitive effects of the proposed acquisition in three areas: (1) the electric home appliance market, (2) the machine shop capability and (3) the metal rolling mill market.
*531In the first area of home appliances, I note that neither of my colleagues apparently disagrees with the district court’s finding of insufficient evidence to establish Allis’ entry into the market, and also, apparently, they concede that the district court had very little data to be impressed by the machine shop capability argument.
The critical area of review, therefore, emerges in the field of metal rolling mills. It is here that the majority part company with the conclusions of the lower court.
In concert, the majority suggests that there is a distinct anti-competitive danger of unlawful reciprocal dealing in the rolling mill-steel industries. My brother Stahl finds an equally anti-competitive effect on the potential entry of Allis-Chalmers into the metal rolling mill market and the potential product extension ramifications of White’s acquisition of Allis’ present capabilities in the field of rolling mill electrical components.
I disagree with these conclusions.
I. Elimination of Potential Competition
In the many points of disagreement between appellant and appellee in the proceedings below, there was one matter about which there was no dispute: it was agreed that Allis could not make out a case of actual competition between the merging companies. Instead, it had to proceed on a theory of “potential competition,” a newly emerging theory which in recent times has been embraced by the courts as a legitimate concern in antitrust considerations.7
Potential, as distinguished from actual, competition must manifest itself in precise forms to come within the scrutiny of those who would enforce the spirit of the Clayton Act. Thus, where two firms market the same or substantially similar products in different parts of the country, their merger might present the question of a “substantial lessening of competition;”8 or where two or more firms merge or form a joint venture for the purpose of entering a market presently dominated by others and it can be shown that but for the agreement each would have entered the new market separately, or that one would have entered independently while the other “remained at the edge of the market, continually threatening to enter;”9 or where because of a long history of repeated entries into a particular market, it was reasonably likely to enter a specific area of that market.10
It becomes obvious therefore that the literal meaning of the word “potential” is not to be applied to the doctrine of potential competition in anti-trust law. He who would urge a case based on this doctrine must perforce prove more than a naked possibility of competition, or that a possible entrant into a market is capable of developing into an actual competitor. It would seem that to forestall a merger because the proposed combination would eliminate potential competition in the anti-trust sense, it is necessary to show first, a market so concentrated that potential competition provides one of the few checks on oligo-polistic pricing and then: (1) that the acquired firm must be a significant factor in that market; and (2) one of the firms, in terms of its objective capabilities must appear to be a likely competitor of the other at some time in the future; or (3) as suggested by some commentators, the potential entrant must appear to be preeminent among the possible likely entrants, or that there are very few others in a similar position.11
*532It is only with these principles in mind that we may appropriately examine the theory of the potential entry of Allis-Chalmers into a new market, potential product extension and potential reciprocity consequences.
This has been done in the discussion hereinafter set forth at length. And having done so, there is no basic area where I find myself in agreement with my brothers. I find the theories advanced on potential entry, product extension, and reciprocity to be devoid of prec-edental support or endorsement by any recognized commentator. Indeed, I view this approach to be an uncharted excursion into a sensitive area of the American economic community, embracing a truly radical concept of the doctrine of potential competition in anti-trust law; a theory which seems to hold that so long as there exists a mere possibility of entry in a market, or a mere possibility of product extension, or a mere possibility of reciprocity, then the proposed merger must be enjoined. I cannot accept this hypothesis; nor, as previously observed, has any previous court decision or commentator suggested it.
I must also regretfully observe that to support this novel theory, it has been necessary to embrace self-serving declarations of opinion by the appellant’s officers as evidence of factual data in critical areas which should have demanded clear analytical proof instead of subjective impressions. Consequently, even were I to accept this unique theory of anti-trust law, I would still be constrained to dissent on the ground that the trial record was deficient in factual support for the theory, heeding always the doctrine that the weight to be given such expressions of opinion is the province of the trier of fact.
II. Reciprocity
Although I agree with the majority’s statement of the law governing anti-competitive effects of reciprocity,12 I find the *533present record devoid of sufficient facts to support a preliminary injunction of the White-Allis acquisition.
From the barest of facts, the majority have conjured vivid overtones of reciprocity in the rolling mill-steel industries. The only statistic capable of being gleaned from the record is that Allis-Chalmers purchases approximately $30,-000,000 of steel annually. To this my brother Stahl adds a figure, not in the record below, of the annual steel purchase of Blaw-Knox; and then a conclusion is proffered, without factual support, that “A White-Allis combination would buy a far larger amount of steel than any of Blaw-Knox’s competitors in the rolling mill market.”13
Even if this were true, which, I emphasize, cannot be determined on the present state of the record, I fail to see how a conclusion of anti-competitive reciprocity follows. Nearly every acquisition has, to some extent, elements of reciprocity. This is especially true where, as here, the industries involved are consumers of the basic component steel.
But the vice of reciprocity, i. e., the ability to transform substantial buying power into a weapon against competitors less favorably situated, can only be determined in the context of the market involved. Otherwise, all that is portrayed is the "potential for reciprocity.” Cf. Consolidated Foods, supra footnote 12, where the company to be acquired, Gentry, along with one competitor controlled 90 per cent of the dehydrated garlic and onion market.
The necessity for detailed information was emphasized in the concurring opinion of Mr. Justice Stewart in FTC v. Consolidated Foods, supra, at 603-605, 85 S.Ct., at 1227-1228:
“Before a merger may be properly outlawed under § 7 on the basis solely of reciprocal buying potentials, the law requires a more closely textured economic analysis. * * * The Act does not require that there be a certainty of anticompetitve effect. But that does not mean that the courts or the Commission can rely on slipshod information confusingly presented and ambiguous in its implications. The law does not require proof that competition certainly will be lessened by the merger. But the record should be clear and convincing that the requisite probability is present.”
The trial record on this point is anything but clear and convincing; it suffers from factual anemia and an examination discloses that it is symptomatic of only a bare suspicion of possible reciprocity. As the Supreme Court has stated, we are concerned with “probabilities, not certainties.” The test is exactly that: probabilities of reciprocity, not possibilities. “We do not go so far to say that any acquisition, no matter how small, violates § 7 if there is a probability of reciprocal buying. Some situations may amount only to de minimis. But * * * here, the acquisition is of a company that commands a substantial share of a market.” Consolidated Foods, 380 U.S. at 600, 85 S.Ct., at 1225.
III. Potential Entry Into the Rolling Mill Market
Based on its review of the evidence, the district court determined:
“ * * * AUis-Chalmers asserts that its entry into the manufacture and con*534struction of metal rolling mills — an area in which Blaw-Knox is active— is sufficiently probable to warrant restraint of White's takeover. However, the only evidence before the Court of Allis-Chalmers potential competition in this field is an assertion by Allis-Chal-mers management that such entry is contemplated. Aside from the fact that Allis-Chalmers now manufactures and constructs paper-making machinery, Tr. 11-12, there is nothing in the record to confirm that Allis-Chalmers has the technological capacity to enter this market. Nor is there evidence to confirm financial capacity to enter, the ability of the market to support an additional competitor, or Allis-Chal-mers’ serious consideration of entry. Finally, the record does not indicate that Allis-Chalmers, by the nature of its present business, stands close enough to the edge of the metal rolling market to exert a competitive influence on others in that industry.”
My brother STAHL concludes, however, that “further inquiry into Allis’ status as a potential entrant into this industry” is warranted. The form of this further inquiry is a discussion of what is described as an “uncontroverted affidavit [that] Allis manufactured components and sub-assemblies for complete metal rolling mills.” An examination of this affidavit, however, reveals only “some information concerning components and sub-assemblies.” And an examination of the source cited in the affidavit, Exhibit QQ, demonstrates that the supportive data itself contradicts the opinion expressed in the affidavit. Far from corroborating the general claim in appellant’s affidavit of “possible manufacture for them of a complete 10 or 12 inch metal rolling mill,” this exhibit merely discloses that Allis-Chalmers sold two heavy-duty tool container block assemblies for use at Alcoa on helicopter blades as a sub-contract with Blaw-Knox and negotiations with Blaw-Knox and not a contract for a $50,000 order for an ingot transfer car assembly for Ford Motor Corporation.14
The only other factor relied upon to buttress the conclusion that “further inquiry” into this area is required, is a vague self-serving declaration by one of the appellants’ officers that Allis-Chal-mers has been secretly negotiating with an unidentified European firm to obtain a license to manufacture an unidentified casting process and unidentified machinery for producing steel slabs. This nebulous statement can hardly be classified as compelling evidence of the probable entry of Allis-Chalmers into the rolling mill market. Furthermore, the analysis ignores such critical determinations as Allis’ questionable financial and technical ability to enter this market, a market which is conceded at another point to possess “already significant barriers to the entry of others.”
*535After examining this sketchy record, self-contradictory in part and notably devoid of persuasive facts, I disagree completely that the trial judge clearly erred in his finding that there was insufficient evidence to establish Allis-Chalmers’ potential entry into the rolling mill market. This is a crucial point. Only by overturning the lower court’s finding of fact that Allis did not intend —or really could not seriously intend with its present capabilities and financial picture — to enter this market can one justify the conclusion that a White takeover could substantially lessen competition in the metal rolling mills market by eliminating Allis as a future potential competitor.15
IV. Product Extension
Stripped of explanatory materials, my brother Stahl’s analysis of “product extension” and “entrenchment” is nothing more than a conclusion that a mere possibility of product extension will probably generate anti-competitive effects outlawed by § 7 of the Clayton Act. It becomes necessary to put this problem in its proper perspective. In the context of anti-trust considerations, product extension is not malum in se. And it is not a factor which has significance in and of itself. It assumes importance only when, by reason of the conduct of the acquired or acquiring party, there occurs such a shift in market power as to produce anti-competitive effects. The law usually is not concerned with ordinary transfers of market power. It is only when the transfer occurs in a market structure so concentrated as to pose a serious threat of creating or enhancing a present monopolistic or oligopolistic market position that it becomes circumspect; and it becomes so because of the probability that competition could be substantially lessened.16
Product extension becomes important in the context of anti-trust considerations only when it becomes a means whereby the proposed merger might enhance market power to the point where it becomes “entrenched.” It is not difficult to predict the probable anti-competitive effect of a takeover by a larger firm of a smaller firm whose product is in a market already concentrated or oligopolistic. Such a merger could increase the existing concentration in the market of the acquired firm by strengthening the position of the smaller, acquired firm.17 Thus, the typical entrenchment case arises in an acquisition of a dominant firm in a relatively small industry by a much larger firm. The market position of the smaller firm may become “entrenched” through predatory pricing, so-called “deep pocket practices” of the parent,18 through various promotional *536advantages not available to the smaller competitors in the market of the acquired company,19 or because of the advantage gained through “product extension”— i. e., the ability of the combined companies “to offer a complete line of equipment to its consumers and to further enhance its position and dominance in the market * * *.”20
In the case at bar we do not have the typical case of a smaller dominant company being acquired by a larger outside company. The record indicates variously that Blaw-Knox controls between 11 and 20 per cent of the metal rolling mills, market. Allis-Chalmers, the acquired company, has none of this market. It is not in the metal rolling mill business. Although it does supply electrical equipment to this market, it is not even a prominent supplier. Suppliers of electrical equipment, with their respective shares of the market, are: General Electric, 45%; Westinghouse, 40% ; Allis-Chalmers, 6%; Reliance Electric, 5%; Cuttler-Hammer, 3%; and Clark Controller (A. O. Smith), 1%.
Although there is an assertion that over the past five years Allis has supplied approximately a half-million dollars worth of such equipment to Blaw-Knox annually, referring to App. 199a for authority, that same source describes these sales in this significant language: “Allis-Chalmers supplies a very small fraction of such [electrical drive devices] components.”
Nonetheless, the conclusion is proffered that in the area of product extension, “the probable anticompetitive effects of an Allis-Chalmers-Blaw-Knox (i. e., White) combination are significant.” This conclusion is based substantially on the interpretation and application of three decisions: FTC v. Procter & Gamble Co., 386 U.S. 568, 87 S.Ct. 1224, 18 L.Ed.2d 303 (1967); General Foods Corp. v. FTC, 386 F.2d 936 (3 Cir. 1967), cert. den. 391 U.S. 919, 88 S.Ct. 1805, 20 L.Ed.2d 657 (1968), and United States v. Wilson Sporting Goods Co., 288 F.Supp. 543 (N.D.Ill.1968).
Judicial application of the product-extension concept to the field of antitrust legislation is far from crystallized. While the relevancy of anti-trust legislation to this area cannot be seriously questioned, this court has already said in General Foods that “we do not read Clorox as proscribing, per se, all mergers labeled ‘product extension mergers’.”
I do not consider the facts in Clorox, General Foods or Wilson Sporting comparable to the case at bar. In all of these cases, the corporations involved, both the acquiring firm and the acquired, were dominant in their respective industries. For example, in Clorox, the acquiring firm, Procter & Gamble, was the giant of the industry, accounting for over 50 per cent of the packaged detergent market and the nation’s largest advertiser of products. In addition, the firm to be acquired, Clorox, was the “leading manufacturer of household liquid bleach, with 48.8% of the national sales.” 386 U.S. at 571, 87 S.Ct., at 1226.
Similarly, in General Foods, this court was confronted with the prospect of General Foods, one of the largest producers of packaged foods in the United States and the third leading national advertiser, merging with S. O. S., the copartner of an industry duopoly, and recognized as occupying “a monopoly position in many areas.”
In the last of the trilogy, Wilson Sporting Goods, the merger involved the ac*537quisition of the “leading manufacturer and seller of gymnasium equipment in the country,” Nissen Corporation, by Wilson Sporting Goods Company, which was the nation’s prominent producer and marketer of sporting goods.
In contrast to the giants represented in these cases, the present appeal involves midgets. If Blaw-Knox controls between 11 to 29 per cent of the rolling mill market, it is by no means the industry’s dominant figure. Allis-Chal-mers has none of it. And even if we are to consider Allis’ share of the market for electrical equipment for rolling mills, its share is only 6 per cent, a feeble percentage when compared to the 45 per cent and 40 per cent share commanded by General Electric and Westinghouse. Nevertheless, from these meager facts, the majority conclude that the probable “anti-competitive effects * * * are significant.”
Upon this emaciated skeleton of facts, the cloak of anti-competitive effects just does not fit. There is no takeover by a larger firm of a smaller one whose product is in a market already oligopolistic; there is no evidence that the market position of the combined firm may become entrenched by predatory pricing or promotional advantages not available to other competitors of the acquired company.
There is only one attempt to inflate the miniscule of facts presented at trial into a full blown Ingersoll-Rwnd argument: because Allis has 6 per cent of the electrical control market, this, when added to Blaw-Knox’s share of the rolling mill market, would produce anti-competitive effects to the prejudice of other rolling mill producers. There was no specific evidence presented at the trial below that the Allis-Blaw-Knox merger would afford advantages not possessed by present rolling mill producers. Other than citing them by name with their respective shares of the market, the record studiously avoids a description of their production and sales capabilities.21 There is no proof that the transfer of market power through product extension advantages will be of such magnitude that Blaw-Knox will become “entrenched.”22
*538V. The Significance of Market Share and Concentration Data
Both my brothers have been careful in their opinions to stress the preliminary nature of this proceeding, as if to intimate that somehow this factor absolves the appellant of demonstrating that the findings of fact were “clearly erroneous” or that the district court erred in construing applicable legal standards under the anti-trust laws.
Conceding that the same quantum of proof was not required at the preliminary injunction stage as would be at the trial on the merits, I suggest that the appellant was obliged to present at least some evidence of relevant facts, figures and statistics describing the affected markets and the shares of these markets commanded by the companies involved. I find it most difficult to justify any reversal of the trial court on the basis of an Ingersoll-Rcmd argument without the presence of such essential factual data.
In this respect, I find it impossible to reconcile a prefatory statement professing not to deal “expressly with the matter of each party’s share of the market in the lines of commerce” with the conclusion that a combined market would “account for a significant share” of that market. This incongruous result is in large measure due to a blend of fact and philosophy which characterize the appellant’s case. Lacking adequate factual support for its thesis of anti-competitive oligopoly, Allis-Chalmers attempted to bottom its case upon a theory of economics, asserting this hypothesis: because White Consolidated is a large corporate conglomerate, its size, without more, will produce anti-competitive effects in affected markets.
This hypothesis is nothing more than a restatement of a “Brandeisian bias in favor of human sized institutions,” a nostalgic attempt to equate bigness with badness.23 But Congress has not written this theory into its anti-trust legislation; nor should any court attempt to legislate such a doctrine into it. So long as the legislative proscription of antitrust activities turns on factors beyond the mere “possibilities” of anti-competitive effects, the courts must be diligent not to substitute the Brandeisian bias for sound analysis.
The characterization of a company as a “large conglomerate” should not impose a presumption of anti-competitive guilt. Section 7 of the Clayton Act nowhere so provides. It is the company’s activities — not its' form and size — which the Congress has sought to regulate. And judicial enforcement of this regulation should only be had upon a factual determination that the activity “may be substantially to lessen competition, or to tend to create a monopoly.” 15 U.S.C.A. § 18.
Moreover, this factual determination should result as the synthesis of an adversary proceeding. It should not be affected in any way by the ex parte actions of the FTC, which in the context of private anti-trust suits is a stranger to the litigation. I reject any inference that the issuance of a proposed FTC complaint is prima facie evidence of illegality, justifying preliminary injunctive relief in this case, and presumably others.
But one other aspect of the majority’s action causes me no little concern. Not only has there been an exercise of trial court prerogatives at this preliminary stage of the litigation, some of the language expressed by my brother Stahl is so sweeping and conclusatory that I fear little remains for an objective determination of the basic issues at the trial still to be had. In substituting philosophy for fact, ex-parte declarations for evidence, such analysis comes dangerously close'to molding a cast for the final result, effectively circumscribing the scope of a case yet to be tried on the merits— despite protestations concerning the preliminary nature of this adjudication. This is amply demonstrated by the fol*539lowing statement contained in the majority opinion:
“Basically, what is at stake in the instant appeal is the life or death of Allis, a viable independent company eager to continue as such, pitted against White, an aggressive, fast-moving acquirer of many diverse businesses, particularly in the past few years.”
First, I question the validity of this conclusion, reached without a trial, that Allis-Chalmers cannot survive as an ongoing business if acquired by White. Moreover, I reject the assumption underlying this statement that the judicial branch has been entrusted with the task of policing the economy, preserving the “viability” of the corporate structure. Such a role cannot be justified within the framework of existing anti-trust legislation. It is only where the “life or death” of a corporate entity has anti-competitive effects that judicial intervention is proper.
The district court concluded that Allis-Chalmers failed to establish a probability of anti-competitive effects. On the basis of the law and the evidence, I do not find that conclusion clearly erroneous.
I would affirm the judgment of the lower court for the reasons set forth in the opinion of the learned trial judge.
Accordingly, I dissent.
. See also Springfield Crusher v. Transcontinental Insurance Co., 372 F.2d 125 (3 Cir. 1967): “There is an abuse of discretion * * * when the action of *528the trial judge is clearly contrary to reason and not justified by the evidence”; New York Asbestos Manufacturing Co. v. Ambler Asbestos Air-Cell Covering Company, 102 F. 890 (3 Cir. 1900), quoted in Murray Hill Restaurant v. Thirteen Twenty One Locust, 98 F.2d 578, 579 (3 Cir. 1938): “The granting of a preliminary injunction is an exercise of a very far reaching power, never to be indulged in except in a ease clearly demanding it; and the decision of a court of first instance, refusing such an injunction, will not, except for very strong reasons, be reversed by this court.”
. “That the trial court could have viewed the facts differently, or that we might perhaps have done so, if we had been the initial trier thereof, does not alone entitle us to reverse. Under Rule 52 (a) and its interpretation in the United States Gypsum Co. case, [United States v. United States Gypsum Co., 333 U.S. 364, 68 S.Ct. 525, 92 L.Ed. 746], there must exist a stronger basis for overthrowing a finding of fact than a mere difference in personal judgment. Such evidentiary weight and such convictional certainty must be present that the appellate court does not feel able to escape the view that the trial court has failed to make a sound survey of or to accord the proper effect to all of the cogent facts, giving due regard, of course, to the trial court’s appraisal of witness credibility where that factor is involved.” Nee v. Linwood Securities Co., 174 F.2d 434, 437 (8 Cir. 1949). See also Judge Learned Hand’s observation in United States v. Aluminum Co. of America, 148 F.2d 416, 433 (2 Cir. 1945): “It is idle to try to define the meaning of the phrase ‘clearly erroneous’; all that can be profitably said is that an appellate court, though it will hesitate less to reverse the finding of a judge than that of an administrative tribunal or of a jury, will nevertheless reverse it most reluctantly and only when well persuaded.”
. I must emphasize in the strongest of terms the extreme preliminary posture of the FTC proceedings. Even if it is proper for this court to take judicial notice of an FTC complaint (much less a “proposed complaint”) made public after the proceedings in the district court, little or no weight should be given to it. A “complaint” does no more than signify that “the Commission * * * [has] reason to believe that [a] person is violating or has violated any of the provisions of sections 13, 14, 18, and 19 of [Title 15].” 15 U.S.C.A. § 21.
The mere filing of such a complaint has little, if any more substantive significance than the filing of a complaint by a private citizen in a civil law suit. Indeed, the filing of the complaint is merely the initial step in the commencement of any action by the FTC against an alleged violator of the anti-trust law. 16 CFR § 3.11. There is no provision for any discovery until after the complaint is filed, and there is no requirement that the complaint be based upon a full-scale investigation of the alleged illegal operation in Part III of the FTC Rules of Practice, governing the nature of adjudicatory proceedings.
The tenor of the FTC Rules of Practice indicate that the bulk of the investigative and adjudicatory function is carried on in much the same manner as in the case of a private law suit. Section 3.12 provides for the filing of an answer, motion for a more definite statement of the charges, admissions, and the administrative equivalent of a default judgment for failure to file an answer. A pre-hearing conference is provided by § 3.21 for the purpose of simplifying and clarifying issues, considering amendments to the pleadings, expediting discovery, issuing subpoenas, etc. Section 3.41 guarantees a public hearing; § 3.42 provides for a presiding official at that hearing; § 3.54 prescribes the method of appealing from the initial decision of the hearing examiner to the Commission.
It should be noted that even after full review by the Commission, the decision is not yet considered final. Only after *530appropriate redress to the courts or waiver of that right as provided by 15 U.S.C.A. § 21 does the Commission’s order become final.
. See footnotes 9, 10, 12, 13, 15, 18, 19, 20, 21, 22, representing ten of the thirty-three footnotes. See also the extensive references and quotations in text of opinion.
. See, e. g., McKesson & Robbins, Inc. v. Charles Pfizer & Co., 235 F.Supp. 743 (E.D.Pa.1964), where it was emphasized that “extensive bearings” on the same issue before the court had been completed by the adjudication section of the FTC after a full evidentiary hearing conducted in an adversary fashion. See also MacIntyre, “Anti-trust Injunctions: A Flexible Private Remedy,” 1966 Duke L.J. 22, 32, wherein the writer asserts that “ * * * the findings of a quasi-judicial administrative agency charged with enforcing the anti-trust laws, made after lengthy contest and after thorough consideration of the evidence, should completely satisfy the burden imposed upon a plaintiff in connection with a motion for provisional relief.” (emphasis supplied)
. Nothing in the schema of the anti-trust statute and the FTO Rules of Practice suggest that the issuance of a complaint or the publication of intent to issue one does in fact reflect the Commission’s views on the merits of the alleged violation. At this stage of the proceedings, the Commission’s task of fulfilling its impartial adjudicatory function remains to be accomplished. Although its prosecutory and adjudicatory functions are not bifurcated by statute, the allocation of duties to different units within the agency effectively separates the two. In addition, the FTC must comply with the requirements of § 5(c) of the Administrative Procedure Act, 5 U.S.C.A. § 1004 (c), in conducting adjudicatory proceedings.
. See Davidow, “Conglomerate Concentration and Section Seven: The Limitations of the Anti-Merger Act,” 68 Colum.L.Rev. 1231, 1241-49 (1968).
. See, e. g., United States v. El Paso Natural Gas Co., 376 U.S. 651, 84 S.Ct. 1044, 12 L.Ed.2d 12 (1964).
. United States v. Pen-Olin Chem. Co., 378 U.S. 158, 170, 84 S.Ct. 1710, 12 L.Ed. 2d 775 (1964).
. FTC v. Procter and Gamble Co., 386 U.S. 568, 87 S.Ct. 1224, 18 L.Ed.2d 303 (1967).
. Davidow, 68 Colum.L.Rev. at 1244. This standard was suggested by Donald *532F. Turner, former head of the Justice Department’s Antitrust Division. See Turner, Conglomerate Mergers and Section 7 of the Clayton Act, 78 Harv.D.Rev. 1313, 1363 (1985). It is to be noted that the instant case is atypical in that the alleged potential entrant is not the acquiring company, as in the usual ease, but the acquired company. This presents a more difficult burden of proof problem: the motivation to prove a future intention to enter and thereby block the merger via the anti-trust laws becomes obvious. Because of this, it is of even greater importance that the alleged potential entrant be judged in terms of its objective capabilities as suggested by Mr. Davidow.
It is one thing for Allis-Chalmers to allege its intention to enter any of a given number of market areas presently occupied by White or one of its subsidiaries, in order to prevent the takeover, but it is quite another to present objective evidence of this intention, and to demonstrate the capability to do so. In this respect, Chief Judge Wright was not content with Allis-Chalmers’ self-serving statement of intention to enter the metal rolling mill area, but insisted upon objective proof of capability to do so:
“However, the only evidence before the Court- of Allis-Chalmers potential competition in this field is an assertion by Allis-Chalmers Management that such entry is contemplated. Aside from the fact that Allis-Chalmers now manufactures and constructs papermaking machinery, Tr. 11-12, there is nothing in the record to confirm that Allis-Chalmers has the technological capacity to enter this market nor is there evidence to confirm financial capacity to enter, the ability of the market to support an additional competitor, or Allis-Chalmers’ serious consideration of entry.” 294 F.Supp. at 1268.
. In FTC v. Consolidated Foods, 380 U.S. 592, 594, 85 S.Ct. 1220, 14 L.Ed.2d 95 (1965), the Supreme Court defined reciprocity as “A threatened withdrawal of orders if products of an affiliate cease being bought, as well as a conditioning of future purchases on the receipt of orders for products of that affiliate * * The Court affirmed the Commission’s decision that § 7 was violated when Consolidated, a giant food wholesaler, acquired Gentry, a smaller company which sold dehydrated onion and garlic to certain of Consolidated’s suppliers. In no uncertain terms, the Court condemned the practice of obtaining a competitive advantage through reciprocal buying practices : “We hold at the outset that the ‘reciprocity’ made possible by such an acquisition is one of the congeries of anticompetitive practices at which the *533antitrust laws are aimed. The practice results in ‘an irrelevant and alien factor’ * * * intruding into the choice among competing products, creating at the least ‘a priority on the business at equal prices.’” 380 U.S. at 594, 85 S.Ct., at 1221-1222.
. The FTC proposed complaint, not in the record, states that Blaw-Knox purchases $42,000,000 of steel per year. Although I consider it inappropriate to consider this item because it is not in the record, it may be considered with another figure, set forth in appellee’s brief, also not in the record: the total annual steel purchases in the United States amounts to 13 billion dollars. Viewed in this light the $30 million purchases of Allis-Chal-mers amounts to % of one per cent of the total; the Blaw-Knox figure would be % of one per cent.
. Reference is made to page 146a of the record. This is an affidavit by Allis’ general manager of paper machinery: “In the custom machine end of the Appleton business, we have manufactured components and sub-assemblies of complete rolling mills. This work has been done for Blaw Knox. Examples of such work for Blaw Knox are shown in Exhibits II, JJ, KK and LL.”
An examination of these supportive exhibits is a fascinating experience. Where one would expect statistics, tabulations, and specialized data concerning “complete metal rolling mills * * * done for Blaw Knox,” these particular exhibits, together with NN, 00, and PP, are nothing more than photographs of pieces of machinery, with no description, no legend, and no identifying data whatsoever!
An examination of Exhibit QQ reveals only this meager information:
“We have an excellent working relationship particularly with Blaw Knox Company of East Chicago which is the Foundry and Mill Machinery Division of White Consolidated Industries.”
“Industrial sale is an integral part of the business of the Paper Machinery Department. The products manufactured for Blaw Knox are especially desirable since they afford optimum utilization of our facilities and equipment in the areas of Foundry, Fabricating, Machinery and Assembly.”
A discussion of the block and car assemblies mentioned above then follows.
. Because I conclude that the lower court’s finding that Allis did not objectively intend to enter the rolling mill market was not clearly erroneous, I do not feel it necessary to detail the reasons for an additional belief that neither of the other two requirements for a § 7 violation based on the elimination of potential competition were satisfied. Suffice it to say : here there was no proof at trial that Blaw-Knox was a significant factor in a market so concentrated that potential competition provided one of the few checks on oligopolistic pricing; that Allis was one of the most likely or most important potential entrants in the metal rolling mills market; or that there were very few others in similar positions at the brink of potential entry.
. See the opinion of Judge Learned Hand in United States v. Aluminum Co. of America, 148 F.2d 416 (2 Cir. 1945). The law of product extension is based on Judge Hand’s thesis that business conduct having the effect of entrenching a monopoly is violative of the anti-trust laws.
. See, e. g., FTC v. Procter & Gamble, 386 U.S. 568, 87 S.Ct. 1224, 18 L.Ed.2d 303 (1967); FTC v. Consolidated Foods, 380 U.S. 592, 85 S.Ct. 1220, 14 L.Ed.2d 95 (1965); General Foods v. FTC, 386 F.2d 936 (3 Cir. 1967), cert. den. 391 U.S. 919, 88 S.Ct. 1805, 20 L.Ed.2d 657 (1968); Ekco Prods., Inc., Trade Reg. Rep. ¶ 16,956 (FTC 1963), aff’d, Ekco Prods. Co. v. FTC, 347 F.2d 745 (7 Cir. 1965); Reynolds Metal Co. v. F.T.C., 114 U.S.App.D.C. 2, 309 F.2d 223 (1962).
. Reynolds Metal Co. v. F.T.C., supra, footnote 17, at p. 229. The essence of this theory is a presumption that the *536“ricli parent” has the financial wherewithal to sell at prices approximating cost or below, and thus to “undercut and ravage the less affluent competition.” Again, this argument obviously has much more forcefulness if the acquiring company fulfills the “rich parent” characterization.
. This theory was suggested in FTC v. Procter & Gamble, supra, footnote 10. The fear here is that the parent, through low-unit cost bulk advertising could cause the acquired firm to become an even more dominant participant in an already concentrated market.
. The best example is probably our own case of United States v. Ingersoll-Rand, 320 F.2d 509, 524 (3 Cir. 1963).
. There was an expression of opinion by one of Allis’ officers — a naked statement with no supportive data — that “should Allis-Chalmers and Blaw Knox be combined, the resulting firm would be the only supplier which could furnish complete-unit responsibility for metal rolling mills and electrical drive and control systems.” 112a. An attempt has been made to equate this self-serving declaration with the force and effect of proved fact. I disagree. Saying it does not make it so.
. These observations on the product extension argument are also applicable to the contentions that there are product extension consequences in the construction equipment and steel castings field.
My brother Stahl’s discussion of construction equipment is most meager. We are not sure what standard is being suggested where, here, as elsewhere, it is stated “ ‘the product extension’ aspects of a White acquisition are deserving of scrutiny in light of Allis’ position as a leading manufacturer of a broad range of construction equipment and its network of dealers who carry a wide range of such equipment.” The trial court has already given this issue its “deserved scrutiny.” It made findings of fact adverse to the appellant, possibly because appellant did not press this point. Factual references to this point on the trial record are conspicuous by their absence. The answer to the fear of product extension in asphalt paving and batching equipment is the same as that proffered in the rolling mills discussion above. The appellant’s own evidence indicates that Blaw-Knox, “in the universe of asphalt plants, pavers and related equipment, [ranks] about sixth with 7 to 8% [of the market share]. With Standard Steel, [a newly acquired company] Allis-Chalmers will rank fifth with 9% in that product category.” I37a.
I see no reason for “further scrutiny.” Facts are facts. It is significant that the cited product extension eases all refer to dominant percentage shares of a concentrated market of the acquired (or combined) company. Blaw-Knox’s 8% share of the asphalt paving market and the Al-lis-Chalmers’ share of the batcher market do not begin to approximate such control, especially in the absence of proof of concentration in this field.
. See Davidow, 28 Colum.L.Rev. at 1285.