These cases concern one of the most important, but one of the most elusive, distinctions embedded in our labor law— the distinction between “primary” and “secondary” union activity under Section 8(b) (4) (B) of the National Labor Relations Act, 29 U.S.C. § 158(b) (4) (B) (1964). J. L. Simmons Company filed charges against Local 742, alleging that the carpenters had violated that section in refusing to install prefabricated doors manufactured by another employer. The trial examiner dismissed the charges. He found that the union and Simmons were the primary parties to a dispute over work preservation and concluded that the union’s pressure on Simmons was therefore permissible. However, the Board reversed, finding that Simmons was a mere “neutral” with respect to the hanging of prefabricated doors and that the union’s actions thus amounted to a forbidden secondary boycott.

The primary-secondary distinction, though originating in the common law, has never taken on very clear dimensions, and the,Board has attempted to bring order out of the chaos by formulating determinate rules to guide its application. One of those rules — the “right to control” test — makes union activity secondary per se if the members’ employer, subjected to pressure, lacks immediate “control” over the matter in dispute. That test was applied by the Board to this case, and its validity is at issue on appeal.1

This court and other courts have recently begun to criticize the “right to control” test. Guided by the Supreme Court’s articulation of the primary-secondary distinction, we have cut back on the test’s application and questioned its *897basic validity.2 Yet the Board continues to defend the test in toto. It takes note of our previous decision undermining the test, but states only that “[w]e respectfully disagree with the Court’s decision, for the reasons set forth in our brief therein, and, accordingly, we shall not burden the Court with a restatement of those views.”3 Fortunately, intervenor J. L. Simmons Company, Inc. has taken a more responsible approach to this appeal. It seeks to save only a part of the “right to control” test by arguing that the test was properly applied on the facts of this particular case which, it says, are significantly different from those in previously reported cases.

While administrative agencies, as a general matter, deserve praise for their efforts to structure discretion under such rules, we hold that in this instance the Board has gone too far. The “right to control” test is not appropriately suited to the statutory purpose it is meant to effectuate. The peculiar facts of this case do not cause us to alter that conclusion. We are convinced that the test, applied as a per se rule, must be entirely abandoned forthwith. Accordingly, we remand this case to the Board for further consideration.

I

Members of Local 742 of the carpenters’ and joiners’ union were employed by the J. L. Simmons Company which had contracted to build a new addition to the Macon County Hospital in Illinois. One of the carpenters’ duties on the project was to hang new doors. This task might have been expected to provide a good deal of work for union members, for it generally involved not simply hanging but also prior preparation — trimming, routing, mortising and cutting— of the doors. The Board adopted a finding of the trial examiner that such preparatory work was “historical and traditional unit work of carpenters at the job-site.” 4 Yet in this case the carpenters were deprived of that part of the task; they were ordered by the Simmons Company to hang premachined doors, prepared at the factory rather than at the job site.

The initial decision to use premachined doors was made not by Simmons but by the Hospital Association through its architect’s specifications. When Simmons originally contracted to build the hospital addition, in July of 1966, the specifications called for wooden doors to be prepared at the job site. But the contract also left the Hospital Association the option to require more expensive prema-chined doors should adequate financing become available. Thus by signing the contract Simmons agreed to install prefabricated doors if asked to do so. The company thereby bound its own hands, *898participating affirmatively in the crucial decision. In August financing did become available and a change order was issued calling for the premachined doors.

Simmons ordered 800 premachined doors from the Anderson Wood Products Company, and three months later 80 of the doors arrived at the construction site. A Simmons vice president, “suspecting that these prefabricated doors were going to become a problem,” contacted the company’s attorney for advice. And a carpenters’ union business agent soon thereafter did raise very strong objections to the deprivation of traditional unit work. “After being apprised of the fact that premaehining was required by the specifications, [he] stated that the [union’s] members employed by Simmons Company would not hang, i. e. install, the doors.” 5 Somewhat more than a week later, the company’s job superintendent requested that a union member begin to hang some of the doors. But the carpenter replied that there was still “a problem” about them and refused.

The union’s ground for refusing to install the premachined doors was that to do so would violate its bargaining agreement with Simmons. There was, however, no specific work preservation clause in the agreement relating to the machining of doors prior to hanging.6 Thus the union had to rely on an implied agreement to preserve work traditionally performed by the carpenters on the job site • — or, at least, an implied agreement to negotiate compensation when prefabricated doors are to be installed.

After the first refusal to hang the doors, the union initiated several attempts to solve the problem through bargaining rather than litigation or continued confrontation. Its attorney got in touch with the company’s attorney and suggested that the doors might be hung if the parties could negotiate a premium to be paid for the work lost. On more than one occasion this opening proposal was made, but the company gave no substantive response and proceeded to file an unfair labor practice charge against the union. Twice thereafter there were requests to hang the doors and similar refusals by the carpenters. Finally, when the company’s unwillingness to negotiate a compromise solution must have been clear, the union gave in and hung all of the premachined doors in the hospital addition.

Both the trial examiner and the Board found that the union business agent’s “threat” to refuse installing the prema-chined doors constituted restraint and coercion within Section 8(b) (4) (ii), and that the initial refusals to hang the doors constituted a “refusal in the course of * * * employment” under Section 8(b) (4) (i). However, the trial examiner found that no unfair labor practice had been committed under Section 8(b) (4). He concluded that the union had not acted with a “secondary” cease-doing-business objective violative of subsection (B). The Board reversed, finding that the carpenters’ actions were indeed prohibited “secondary” pressure against Simmons. It relied solely on its “right to control” test. The one crucial factor, it said, was that Simmons had no control over the type of doors to be installed and thus over the substance of the union’s grievance. In its formal opinion, the Board considered no other factors whatever in reaching its result. It concluded that simply because “control” lay with the hospital the “Simmons Company was a neutral with respect to the assign*899ment of this work. Hence, the pressure exerted against Simmons Company, the secondary employer, was for the purpose of forcing it to cease doing business with the Hospital and forcing the Hospital to cease doing business with Anderson Wood Products Company for the furnishing of premaehined * * * doors.”7

II

Only eight months ago, this court held in a similar case that exclusive reliance on the “right to control” test was impermissible. Local Union No. 636, United Assn of Journeymen & Apprentices of Plumbing & Pipefitting Industry v. NLRB, 139 U.S.App.D.C. 165, 430 F.2d 906 (1970). In that decision we directly followed our two other sister circuits which recently have spoken on the issue. American Boiler Manufacturers Assn v. NLRB, 8 Cir., 404 F.2d 556, 560-561 (1968); NLRB v. Local Union No. 164, Int. Brotherhood of Electrical Workers, 3 Cir., 388 F.2d 105, 109 (1968). In all of these cases, the union exerted pressure on its members’ employer to regain unit work of which they had been deprived. In all of these cases, the Board had given dispositive weight to the fact that the employer was contractually bound to use prefabricated products. The only difference between this case and the previous cases is that here the union could invoke no specific work preservation clause in its agreement with the employer. The problem is whether this difference makes a difference. We conclude that it does not.

The defect of the per se “right to control” test is at its very core, its theoretical foundation. The test rests upon a narrow, mechanical — and mistaken — conception of the primary-secondary distinction. To label an employer as a “neutral” in a labor dispute involving his own employees, the test does not look to the substance of the dispute, to its history, or to its apparent motivation. It ignores all of these factors. And, instead, it focuses solely on whether the employer is immediately capable of granting the union what it wants. If the employer is not capable of so doing, he is labelled a “neutral.”

The artificiality of the test is particularly patent in cases such as this one— where the employees apparently have a grievance growing out of labor relations with their own employer (i. e., loss of traditional unit work) and immediately provoked by an action of their own employer (i. e., his signing of a contract which would deprive them of that work).8 The Board does not stop to consider these aspects of the dispute. Nor does it look to see whether there may be other grievances, against other employers, which underlie the union’s action. Rather, the Board seems to hold that, when an employer contracts away his power to satisfy union demands, he also contracts away his interest and involvement in the labor relations issue in dispute. He is depicted as the mere passive agent of others (¿. e., the party who required the prefabricated products and/or the party who manufactures them); the fact that the present dispute has its origin and its substantive focus in his ongoing labor relations with his employees is deemed irrelevant. This Cinderella-like transformation of an obviously involved party into a neutral bystander is rather incredible.

The legal effect of the Board’s test is to allow an employer to bind his own hands and thereby immunize himself from union pressure occasioned by his employees’ loss of work. In one act, the employer helps to create a labor conflict and simultaneously wash his hands of it. The result goes far beyond the purpose of prohibiting secondary union activity: to limit the arena of economic conflict, to protect uninvolved, truly neutral par*900ties from becoming involved in labor disputes not their own.9 Section 8(b) (4) (B), specifically limited to secondary activity, is not an escape hatch for an obviously involved employer. Its interpretation requires a more complete and more realistic conception of involvement in a labor dispute than the Board now provides. The substance, history and motivation of a particular dispute must not be ignored.

The importance of these substantive factors to the definition of secondary union activity has been authoritatively established at least since the Supreme Court’s decision in National Woodwork Manufacturers Assn v. NLRB, 386 U.S. 612, 87 S.Ct. 1250, 18 L.Ed.2d 357 (1967).10 That case, like this one, concerned a dispute over work preservation, although there the dispute centered on a specific work preservation clause in the bargaining agreement. The Court spoke generally to the meaning of “secondary” activity under both Section 8(b) (4) (B) and Section 8(e). It prescribed the sort of inquiry which must be used to determine whether the pressured employer is but a neutral dragged into a dispute not his own:

“an inquiry into whether, under all the surrounding circumstances, the Union’s objective was preservation of work for [unit] employees, or whether the agreements and boycott were tactically calculated to satisfy union objectives elsewhere. Were the latter the case * * * the boycotted employer would be a neutral bystander, and the agreement or boycott would, within the intent of Congress, become secondary. * * * The touchstone is whether the agreement or its maintenance is addressed to the labor relations of the contracting employer vis-á-vis his own employees. * * ”

386 U.S. at 644-645, 87 S.Ct. at 1268-1269. (Footnotes omitted; emphasis added.)

The National Woodwork Court did not apply the above primary-secondary reasoning to the “right to control” test, since the issue had not been presented.11 But it is quite clear that the two are in fundamental conflict. The Board’s test flies in the face of National Woodwork in three respects: (1) It elevates one circumstance — where the immediate “control” lies — to per se status, rather than evaluating “all the surrounding circumstances.” More than mere form is at stake here. The Supreme Court properly recognized that realistic assessment of a union’s objective in a particular situation is a complex, subtle matter and must depend on a variety of evidential factors. A mechanical per se rule simply cannot sift and weigh the evidence with the required sensitivity. (2) A more basic failing of the “right to control” test under National Woodwork is that it focuses on entirely the wrong set of circumstances. It is concerned solely with which party presently has the *901power to satisfy the union’s objective, rather than focusing on the substance of the objective itself.12 Thus it misses the point of the primary-secondary distinction as set forth in National Woodwork. (3) And, more particularly, the Board’s test fails to evaluate the substance of the union’s objective by looking to see whose labor relations it is addressed to. That is the crucial matter under National Woodwork. If the union’s grievance has to do with some third party’s relations with his employees, the pressure it exerts against its members’ employer may well be secondary.13 But National Woodwork makes clear that, if the grievance has to do with the labor relations of the pressured employer with his own employees who are exerting the pressure, then the activity is probably primary and permissible under Section 8(b) (4) (B). The Supreme Court also held that an objective of preserving traditional work for the union’s members with the pressured employer is definitely primary. The fact that the employer may have to cease doing business with another party to satisfy the union’s demand is a permissible ancillary effect.14 Of course, there may be cases where the union’s objective is not clearly, or only partly, to preserve its members’ work. But the per se “right to control” test would declare an activity to be secondary without any inquiry whatever into such matters.

Each of the circuit decisions, supra, rejecting the Board’s test has found the test to be irreconcilable with the primary-secondary analysis of National Woodwork. In Local Union No. 686, supra, we stated that “the Board’s application of the ‘right to control’ test completely ignores the teaching of the Supreme Court in National Woodwork * * 139 U.S.App.D.C. at 168, 430 F.2d at 909. The above discussion in this opinion is designed to reinforce that conclusion. We concluded before that “the ‘right to control’ test must be abandoned.” 139 U.S.App.D.C. at 169, 430 F.2d at 910. We do not retreat from that position.

Because the defect of the “right to control” test lies at its very core, the particular fact situations to which it is applied are inconsequential to its validity. It may be useful, however, to discuss briefly the extent to which the facts of this case vary from those that went before. In addition to the dispositive discussion of National Woodwork, our opinion in Local Union No. 686 analyzed more narrowly the equities in the particular case. We emphasized the fact that the employer theré had signed a specific work preservation clause in his labor agreement and that he could have avoided the labor conflict by respecting that agreement. His first duty, we said, was to honor a contract already signed with *902the union rather than to hold to a conflicting contract later signed with a third party.

This narrow line of reasoning, of course, does not precisely apply to the facts of the case now before us. There was no specific work preservation clause in the bargaining agreement here. Thus the intervenor, J. L. Simmons Company, argues that it is unfair on these facts to allow the union to exert the pressure that it did. However, continuing for the moment in the balance-of-equities framework, we may note that Simmons was by no means caught by surprise when a problem developed over installation of prefabricated doors. Indeed, the trial examiner found that a Simmons vice president anticipated the dispute and sought legal advice even before the union made its objection known. Since the preparation of doors on the job site was historical and traditional unit work, Simmons surely had notice of the impending conflict when it signed its contract with the Hospital Association.

It is true that there is no question here of honoring the letter of one contract over another, since the labor agreement was apparently silent on the work preservation issue. But it is an established and fundamental principle of labor law that labor-management contracts are not to be read with narrow precision. In the words of the Supreme Court, they call “into being a new common law — the common law of a particular industry or of a particular plant.” United Steelworkers of America v. Warrior & Gulf Navigation Co., 363 U.S. 574, 579, 80 S.Ct. 1347, 1351, 4 L.Ed.2d 1409 (1960). “‘One cannot reduce all the rules governing a community like an industrial plant to fifteen or even fifty pages.’ ” Ibid. “Gaps may be left to be filled in by reference to the practices of the particular industry and of the various shops covered by the agreement.” Id. at 580, 80 S.Ct. at 1352. Thus the silence of Simmons’ labor agreement as to work preservation does not take that agreement out of the picture. If the “practices of the particular industry” may be read into the common law of the agreement, then the preservation of traditional unit work has substantial status under Simmons’ contract with the carpenters’ union.15 Not only the Hospital Association, but also the union, had a contractual claim on Simmons. And the union’s claim was the prior one.

We do not intend to hold that preservation of the door-preparation work was necessarily to be implied into the labor agreement. But we do believe that it had an important status under the agreement. If labor peace is the overriding goal of our labor laws, it seems clear that peace would have been best preserved here if Simmons, after unilaterally signing a contract which it should have known would deprive its employees of work traditionally theirs, had been willing to stop and bargain with the union over the prefabricated door issue, rather than refusing to hear of any compromise. See Fibreboard Paper Products Corp. v. NLRB, 379 U.S. 203, 85 S.Ct. 398, 13 L.Ed.2d 233 (1964).16 Even after creating the conflict, Simmons was, after all, not completely without resources to work out a solution to the dispute. Some financial *903compensation, for example, could have been paid to the employees in return for hanging the prefabricated doors. In the interest of labor peace, the union also would have been best advised to forego all “threats” and to stick exclusively to rational bargaining.

Finally, however, we must emphasize that this discussion of the equities of the situation is not dispositive of this case. The dispositive consideration — and the consideration which dictates the abandonment of the per se “right to control” test — is the process of inquiry prescribed in National Woodwork, supra, to govern the primary-secondary question. Neither the Board’s view nor our own view of what is, broadly, fair or unfair may be put into effect using the vehicle of the Section 8(b) (4) (B) prohibition of secondary pressure. See NLRB v. Insurance Agents’ International Union, 361 U.S. 477, 497-498, 80 S.Ct. 419, 4 L.Ed.2d 454 (1960). If the presence or absence of a specific work preservation agreement is to make a difference, it can only be as one of many “surrounding circumstances” used to determine what the union’s objective was. It may be, for example, that when the union acts to enforce a specific clause, its objective more clearly relates to its labor relations with the primary employer. But the absence of such a specific clause is obviously not enough by itself to indicate a secondary objective relating to another employer. We note that the focus of National Woodwork is whether the union’s objective was directed to the labor relations — not simply the specific contractual provisions, but the labor relations generally — of the union members’ employer. There is, then, no bright line protecting employers such as the Simmons Company. Full immunization of an employer who desires to contract with a third party to install prefabricated products must be left to future bargaining agreements or future congressional mandates.

Hence we remand this case to the Board for a new decision, considering the union’s objective “under all the surrounding circumstances.” Of course, if it throws any light on the union’s objective, the Board may consider Simmons’ lack of “control” as one of those circumstances. It may also consider the absence of a specific work preservation clause to the extent that that bears on the substance of the union’s objective. We emphasize once more, however, that as a per se rule the “right to control” test must be finally abandoned.

Remanded for further proceedings.

I concur in the foregoing result and in the opinion, however, I wish to state my view that the mere absence of a work preservation clause in a collective bargaining agreement is not a sufficient fact to authorize the Board to completely dispose of the case by use of a per se “right to control” test. Further, that nothing we have said concerning the absence of a work preservation clause should be construed as indicating our opinion as to the inferences to be drawn from that fact when the Board, on remand, considers all the circumstances of the case. At that time the Board should consider all the facts and come to its own conclusion after considering all possible tests of the union’s objectives.

(concurring specially):

I agree with Judge Wright’s conclusions as to the “right to control” test and concur in the remand.

I think, however, that the Board should initially interpret the Collective Bargaining Agreement and I do not join in that part of the opinion which deals with the interpretive problems.