TABLE OF CONTENTS
PAGE
I. INTRODUCTION 335
II. FINDINGS 339
A. The Hiring System 339
B. The Philadelphia Plan 342
C. The Benjamin Franklin Programs I and II 345
D. Procedural History 347
E. The Delays and Contempt Issues Involving Abraham Freedman, Esquire 348
F. The Named Plaintiffs as of Certification 350
G. Statistical Evidence 350
1. Membership Disparities 351
2. Discrimination in Entry 352
3. Disparities in Hours and Wages 353
4. Referrals 355
H. Other Proof 357
1. Entry Discrimination 358
2. Individual Testimony 360
(a) Samuel Long 360
(b) Willis Fox 360
(c) Robert Ahmad 361
(d) John Dent 362
(e) Elijah Dukes 363
*334(f) Charles Iseley 363
(g) John Dodson 364
(h) Lloyd Hudson 364
(i) Duáne B. Johnson 365
(j) George Benjamin 366
(k) Timothy A. Roundtree 367
(l) Cleveland Allen 368
(m) Conclusion 369
I. Rebuttal 369
1. Experts’ Statistical Analysis 370
a. Labor Pool 370
b. Features of Entry 375
c. Hours and Wages 377
i. Dr. Wachter 377
ii. Dr. Perl 378
iii. Dr. Dempster ■ 379
d. Conclusion 380
2. Other Rebuttal of Defendants 380
J. The Case Against JATC 381
K. Glasgow, Inc. and the Associations 384
III. LEGAL CONCLUSIONS 386
A. Appropriateness of Plaintiffs’ Class: Considerations of Standing and the Requirements of Rule 23 386
1. Adequacy and Standing 387
2. Rule 23(b) 389
3. Subclassing 390
4. Motion for Amendment of Plaintiffs’ Class Definition 391
B. Substantive Claims 394
1. Claims Against the Union 394
a. Title VII 394
i. Title VII Jurisdictional Issues 394
(a) Conciliation 394
(b) Scope of Suit 395
ii. The Merits of the Title VII Claim 397
(a) Intentional Discrimination 398
(b) Disparate Impact Discrimination 399
b. Section 1981 Claims Against Union 399
c. Section 1985(3) 401
2. Claims Against Associations and Contractors 401
a. Section 1981 401
1. The NLRA Cases 402
2. Civil Rights Cases Rejecting Vicarious Employer Liability 403
*3353. Cases In Which an Employer or Employer’s Association Was Held Liable for Discrimination Arising From Terms of Contract 407
4. Doctrine of Respondeat Superior 409
(a) Analogies to Civil Rights Suits Against Municipalities or Supervisory Personnel 409
(b) The Application of the Doctrine of Respondeat Superior 411
b. Employers and Associations Potential Liability Under § 1985(3) 413
C. Appropriateness of the Defendant Class as to Section 1981 Claims 414
1. Rule 23(a) 414
2. Rule 23(b) 415
3. Standing 417
4. Personal Jurisdiction Over the Defendant Class 419
OPINION
A. LEON HIGGINBOTHAM, Jr., Circuit Judge.*I. INTRODUCTION
This employment discrimination suit was instituted in 1971 by twelve black plaintiffs on behalf of a class of minority workers involved in or desiring admittance to the operating engineer trade in Eastern Pennsylvania and Delaware. Also a named plaintiff is the Commonwealth of Pennsylvania, suing on behalf of its citizens and the above-described class. Defendants in this action are as follows: Local 542 of the International Union of Operating Engineers; a class assertedly represented by Glasgow, Inc., consisting potentially of more than 1400 construction contractors and employers receiving referrals through Local 542’s exclusive hiring hall; four construction trade associations which represent the employers in contract negotiations;1 and the Joint Apprenticeship Training Committee (“JATC”), an organization created by Local 542 and the trade associations for the induction of new operating engineers. This opinion constitutes the findings of fact and conclusions of law relevant to the liability stage of the trial of this massive and intricate case.
The active claims of plaintiffs’ class include a Title VII, 42 U.S.C. § 2000e et seq., employment discrimination claim against all defendants based upon alleged discrimination in the membership practices of 542, the operation of its referral system, and the hours and wages of minority operating engineers. The foundation of this claim also serves as the basis for a 42 U.S.C. § 1981 claim and a conspiracy claim under 42 U.S.C. § 1985(3) against all defendants, an unfair representation claim under 29 U.S.C. § 158 against Local 542, and a Title VI discrimination claim against all those of the defendants who have deprived minority in*336dividuals of the benefits of federally funded construction projects. Local 542’s function as an exclusive hiring hall in its geographical jurisdiction is at the center of each of the above claims; however, plaintiff asserts that the contractors and associations, having agreed to such a system in 1961 or thereafter, are co-participants with the union and cannot be absolved from liability for discrimination in the operation of the hiring hall. Although plaintiffs have frequently characterized their action as being based on intentional discrimination they are not limited to this standard particularly in view of the broad allegations in the complaint. I must therefore consider plaintiffs’ factual claims in light of the full range of potential liability under the civil rights statutes involved.
Of course, a very significant aspect of the instant suit is its class action status. As originally certified on March 13,1972, plaintiffs’ class was divided into the following subclass descriptions:
(a) all minority group members who currently have the skills, when measured by objective standards, of at least a journeyman operating engineer and who work, or may work, within the territorial jurisdiction of defendant Local 542;
(b) all minority group members who are partially skilled, when measured by objective standards, to perform operating engineers work and who work, or may work, within the territorial jurisdiction;
(c) all unskilled minority group members who wish, or may wish, to acquire skills in the operating engineers trade and who are physically capable of acquiring such skills and performing operating engineers work within the territorial jurisdiction of defendant Local 542.
Among a number of threshold issues in this case, I also have before me a request by plaintiffs to recertify plaintiffs’ class as well as arguments against certification of plaintiffs’ class as now constituted and as proposed by plaintiffs. These arguments by defendants, and several unnamed members of the defendant class which this court has permitted to appear post-trial as “non-parties,” 2 assert, inter alia, the inadequacy of named plaintiffs as class representatives, the lack of standing of named plaintiffs, and the non-commonality and atypicality of issues, claims and defenses.
Other threshold issues before me now are plaintiffs’ request for recertification of the defendant class so as to comport with the applicable statutes of limitations, and the request of non-parties for decertification of the defendant class of contractors. As originally certified on March 13, 1972, that defendant class now consists of:
(a) all contractor associations which are, or may be, parties to a collective bargaining agreement with Local 542, International Union of Operating Engineers; and
(b) all contractor-employers who are subject to collective bargaining agreements with Local 542, International Union of Operating Engineers, and who, pursuant to such agreements, employ or will employ operating engineers referred to them by defendant Local 542.
There are an assortment of objections to certification of the defendant class including the claimed lack of personal jurisdiction over the unnamed class defendants, plaintiffs’ lack of standing to assert claims against a defendant class and the asserted inappropriateness of defendant class certification under rule 23(b)(2), F.R.Civ.P.
For the reasons expressed below I hold that this action is maintainable as both a plaintiff and defendant class action, given the modifications as will be set forth. Furthermore, we find the defendant 542, JATC, and the defendant class and associations liable injunctively.
The issue of individual monetary recovery as well as the possible issue of class compo*337sition for purposes of damage relief should not now be decided but must await Stage II of this litigation in accordance with this court’s bifurcation order. The issue of damages, if any, owed to the plaintiffs who initiated this suit or to any members of the purported class involves a separate evidentiary inquiry and a further legal analysis which would not be appropriate at this stage. The case was purposely bifurcated with all parties recognizing that in a variety of ways, upon a determination of liability, the discovery and trial of the damage issues could be expedited. It would be unnecessarily costly to have expanded and made this litigation even more protracted by considering the damage issues prior to the time when the correctness of my present findings and judgment on the liability issues have been fully subjected to final appellate review. After a final decree has been entered on the liability issues, I will grant a petition under 28 U.S.C. § 1292(b) certifying that an immediate appeal from the order “may materially advance the ultimate termination of the litigation . . .”
The facts of the instant case, as detailed below, demonstrate the complexity and subtlety of the interrelationship of race, collective bargaining, craft unions, the employment process and that ultimate goal — real jobs which offer adequate hourly compensation and reasonably consistent pay checks through the year. Here there are many contradictions between pronounced policies and actual practices. Also there are some aspects of viral nepotism at its worst which had a disproportionate impact against blacks but also affected many whites. Some of the practices cannot be categorized as exclusively beneficial to all whites or as exclusively harmful to all blacks. Thus there has to be a careful weighing of the relative racial impacts of many practices and policies.
Here we have not been confronted with policies which announce publicly doctrines of racial exclusion or segregation as has occurred in some cases in the nation’s past where such doctrines were announced either by legislative fiat or by proclaimed union or corporate policy. See Griggs v. Duke Power Company, 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971); Brown v. Board of Education, 347 U.S. 483, 74 S.Ct. 686, 98 L.Ed. 873 (1954); Derrick A. Bell, Jr., Race, Racism and American Law (1973); Sterling D. Spero and Abram L. Harris, The Black Worker (1931, reprint ed. 1968); Herbert R. Northrup, Organized Labor and the Negro (1944); Robert C. Weaver, Negro Labor (1946); H. R. Northrup et al., Negro Employment in Basic Industry (1970); Negro Employment in Finance (1970), vol. 2; Negro Employment in Public Utilities (1970), vol. 3; Negro Employment in Southern Industry (1970), vol. 4; Negro Employment in Land and Air Transportation (1971), vol. 5. See also House Committee on Education and Labor, H.R.Rep. No. 718, 89th Cong. 1st sess. 1965; Herbert Hill, “Racial Inequality in Employment: The Patterns of Discrimination,” THE ANNALS 357 (January 1965), pp. 30-47. In some instances some of the union members seemed concerned about increasing the opportunities for minority members; yet more often than not there was discrimination on a more sophisticated and subtle level, even though the consequences could be almost as devastating as the most crude form of discrimination. At the critical level of viable jobs and equal opportunities, there were intentional and persistent efforts to exclude and discourage most of the minorities who, but for their race, would have been considered for entry into the union and for the more lucrative jobs.
Of course labor unions and the collective bargaining process are part of the American democracy. This process has improved dramatically the options of many. As Professor Archibald Cox has observed:
The purpose and effect of every labor organization is to eliminate competition in the labor market. Chief Justice Taft’s classic statement observed:
“[Labor unions] were organized out of the necessity of the situation. A single employee was helpless in dealing with an employer. He was dependent ordinarily on his daily wage for the mainte*338nance of himself and family. If the employer refused to pay him the wages that he thought fair, he was nevertheless unable to leave the employ and to resist arbitrary and unfair treatment. Union was essential to give laborers an opportunity to deal in equality with their employer.”
Each bricklayer’s local seeks to control the supply of bricklayers’ services available to contractors within its geographical jurisdiction. United Steelworkers of America controls the supply of labor available to United States Steel Corporation. In this sense every union is an avowed monopolist.
Cox, Labor and the Antitrust Laws — A Preliminary Analysis, 104 U.Pa.L.Rev. 252, 254 (1955) (footnote omitted).
Similarly, after quoting the above passage, Judge Aldisert has recently observed that:
“the very essence of the labor movement, as protected by the national labor policy, hinges on labor’s ability to seek monopoly in appropriate spheres . . . .”
Muko v. Southwestern Pennsylvania Builders and Construction Trades Council, No. 75-979, slip op. at 6 (3d Cir., August 11, 1978).
Yet it is clear that by the nation’s civil rights laws there are limits to which labor’s “avowed monopolistic” powers can be extended. Here the union (with the involvement of the contractors) has gone beyond the “appropriate spheres” of national labor policy and consequently has breached the overriding civil rights laws. Tragically, blacks and other minorities seeking entry most often were not aided but instead were deterred by willful manipulation of the collective bargaining process. Measured by the actual results, the antidiscriminatory proclamations have proven to have been mere rhetoric. During his campaign for election as business manager Robert Walsh’s platform was “fair play.” Yet, during his administration viable, equal job options were most often denied to minorities, and those relatively few minorities who received the option were seldom granted the more lucrative long term jobs.
Many explanations are proffered for this despicable state of affairs. Some assert that the racial disparities were caused merely by nepotism or union politics. Yet, while it may seem commendable for fathers to pass on to their sons and to their other relatives a heritage of lucrative employment, a union of elected officials all of one race cannot use their power to implant racially chauvinistic and discriminatory systems and union power politics create no special immunity from civil rights law enforcement. With intensity some employers urge that they agreed to the exclusive hiring hall system solely as a matter of economic survival at the end of a destructive ten week strike when the union would not compromise for any other hiring alternative. Yet economic pressures, however strong and harmful they might be, do not create immunity for employers, at least not in this liability phase.
Finally, it must be emphasized that the economic stakes here are high. Members of the union could earn from $5.93 to $12.35 per hour. Some earned more than $30,000 and a few more than $40,000 per year. By now the average wage for the actively employed member is in excess of $15,000 per year. Despite its monopolistic powers, no one union is required to open its doors fully to admit everyone who applies — for to do so would reduce the financial options of the more senior union members. Nevertheless, a union cannot manipulate the collective bargaining system in claiming that it is restricting entry options to protect the general economic interest of its members while at the same time opening the back doors manipulatively to permit its special friends, relatives and others to enter through a racially discriminatory system. In Local 542, there was extraordinary manipulation of the process by many individuals who maneuvered the system for their relatives or their special friends — most of whom were white. Black veterans who had served their country with honor and distinction by proficiently operating engineering equipment during the Vietnam War and World *339War II were dissuaded persistently from applying to the union. Yet the young sons of the business manager Robert Walsh could enter the union with ease: one of Walsh’s sons entered at the age of 18 and by his second year was earning more than $43,000 per year. The son of the hiring hall dispatcher at the age of 18 entered without any prior experience as an operating engineer’s apprentice or oiler and in violation of the union rules. This pattern of easy, inappropriate entry and assignment to better jobs was repeated time and time again for the special friends of the union. See generally Plaintiffs’ Schedule On Union Nepotism, Volumes, I, II and III. Union officials maneuvered the system so that their friends and relatives achieved entry and better jobs while those on the outside of the union power structure — including both blacks and whites — were deterred, delayed or refused union membership and access to the hiring hall. Though a relatively few blacks benefited, the totality of the union’s conduct demonstrates that it sought to aid a selfish cause in part by a deliberate policy of resistance to equal minority participation as is shown by the deceptions contained in the defendants’ 1968 Affirmative Action Agreement, discussed below, and by a series of other specific acts which precluded minorities, mostly blacks, from getting the options they otherwise might have had.
II. FINDINGS
A. The Hiring System
An examination of the facts in this case must begin with an elaboration upon the structure and operation of Local 542.
In 1961 Local 542 concluded a bargaining agreement with the defendants Contractors Association of Eastern Pennsylvania (“CAEP”) and General Building Contractors Association (“GBCA”) according to which 542 became the exclusive mechanism through which operating engineers were to be employed. Another association, the Delaware Construction Association, had agreed to the exclusive hiring hall earlier in 1961. Prior to 1961 the negotiating trade associations had resisted efforts to establish an exclusive hiring hall. Given the reality of a ten week strike by 542 members and the prospect of serious adverse economic consequences from a continuing strike, the CAEP and GBCA agreed to the union’s proposal.3
The hiring hall referral system thus formulated on paper is essentially the same one as is now embodied in the collective bargaining agreement. (I am not suggesting that the hiring hall in fact operated as it was supposed to or that hiring classifications were uniformly honored.) By the terms of the hiring hall agreement 542 is to maintain lists of operating engineers, or would-be engineers, in four basic categories (“groups”) which are defined by measuring hours of recent construction experience. When an employer has need of an operating engineer he will notify 542 and within 24 hours should receive a referral. The key provisions of the Bargaining Agreement (“agreement”) are as follows:4
GROUP I
GROUP I shall consist of all those applicants who are under Group I and who have qualified for same as of May 1,1963, and all applicants who have worked within the geographical area of the Eastern half of the State of Pennsylvania and the State of Delaware five thousand (5,000) hours within the past eight (8) years; or two thousand five hundred (2,500) hours for the past three (3) years. All applicants shall be entitled to priority under Group I who would have qualified for any of the foregoing classifications were it not for absence due to military service, or service as a Union official, or disability within the provisions of Local 542 Operating Engineers Welfare Plan,
*340GROUP I-A
[This classification added in 1968 by amendment]
GROUP 1-A — Shall be reserved for such jobs that are not physically demanding. This normally would be equipment such as compressors, welding machines, generators, heaters, etc. But recognizing that such or similar jobs may be physically demanding in certain instances, the Employer will discuss the details of such or similar jobs with the Union in order to place the job in the proper category.
Registration on Group I — A out-of-work list will be voluntary and will be restricted to those who are fifty (50) years of age or over, and who have been continuously in Group I for a period of fifteen (15) years of more, or those who are physically handicapped and who are unable to pursue their normal occupational classification. .
GROUP II
GROUP II shall consist of all those applicants who qualify for any of the following categories: All applicants who have worked within the geographical area of the Eastern half of the State of Pennsylvania and the State of Delaware twenty-five hundred (2,500) hours within the past twelve (12) years; four hundred (400) hours per year during any three of the past five years; or four hundred (400) hours during the past year.
GROUP III
GROUP III shall consist of all other applicants for employment. [Article II, § 2(e).]
According to Article II, § 2(f), the union is to maintain.a separate list for each group. One’s position on the list depends on the date of his registration certifying that he is available for work. Under section g the contract provides that the Group I list is to have priority. After Group I listees have been placed, Groups I — A, II, and III are to be used. The agreement provides one basic exception from the requirement to refer according to date of registration and that is with respect to jobs requiring special skills. If an employer needs a specially skilled person, such as a bulldozer operator or a crane operator, the hiring hall is empowered to select such a person even if it means bypassing persons on the list ahead of him not possessing the requisite skills.5
Although the hiring hall constitutes the exclusive referral system, each employer has the right, based on his determination of competency, to refuse to employ one who is referred. If an employer makes such a refusal he may receive another referral which in turn he can accept or refuse. Under Art. II, § d, however, selection for referral shall be made without discrimination.6
As originally constituted, for all purposes of this suit, the divisions of 542 consisted of the parent body, the A and B branches, and the C and D branches. A Registered Apprenticeship Program was also instituted in 1966. The parent body was to consist of experienced operating engineers. The A branch members were to be unskilled oilers, beginners in the trade; and B branch members were to be the operators of earth-mov*341ing equipment. The C branch members were to work in yards and shops and D branch members were surveyors: neither of these two branches were directly subject to the hiring hall system negotiated and agreed to by the contractor associations. Registered Apprentice (RA) entrants were to include those novices, frequently doing unskilled work as oilers of machinery, who seek entry into the union’s construction branches. This entry could be achieved upon attaining the status of journeyman operator.
Until 1972, A, B, or C branch members could not run for or hold union office, and only parent body men7 were eligible for appointment as master mechanics, a supervisory position usually paying relatively high rates. In 1973, however, the A and B branches were eliminated. Their members were transferred to the parent body. The D branch was supposedly eliminated by the end of 1975; however, according to one union exhibit its existence continued at least until January 2, 1976.8 At the present time the parent body, C branch and the RA program are the only divisions of 542.
The registrant program is a classification by which an operating engineer not eligible for union membership maintains his referral status. This classification was created in 1965 and was initially divided into A (inexperienced) and B (experienced) registrants. That distinction was theoretically phased out after 1968 because the A registrants were detracting from the work available to registered apprentices. Since 1970 a field admission test and a written test have been required of applicants for registrant status. Registrants can progress to Group I status without seeking to join the union. (No evidence has been presented concerning the precise number of registrants who might not seek union membership status.) After achieving 2500 hours, however, a registrant can, under the contract, be admitted into the union upon application.
Formal entry into the union may also be achieved, according to procedure established by Local 542, as detailed below, through union organization of a workforce of an employer not previously subject to a bargaining agreement with Local 542. Prior to January 1, 1975, construction employees who were organized might enter either A or B branch. Since January 1, 1975 when these branches were abolished, all such employees are to enter into the parent body directly. Yard and shop workers who are organized are to enter C branch and surveyors are to enter the D branch. Intra-union transfers from A and B branches to the parent branch after four and three years respectively are and have,been available. Transfers out of C branch have occurred either without restriction or after one year of experience in C branch.9
Geographically, Local 542 encompasses Delaware and Eastern Pennsylvania. Because of the enormous size of this jurisdiction, the Local is divided into five districts, each with its own referral lists and hiring hall but all ultimately under one administration. It appears that the normal but not necessarily the absolute practice is for the work site hiring hall to make referrals based on its own district’s listings.
The estimated 1400 contractors who have recently engaged in operating engineering work within Local 542’s wide range may vary considerably in size. Relatively few are members of the defendant associations.
CAEP’s active members numbered only 109 during the period 1965-1971, 88 during 1972-73, and 83 during 1974-75. Defendant GBCA listed its active membership at 104 for the 1965-1971 period, 99 for 1972— 73, and 83 and 74 for 1974 and 1975, respectively. Defendant United Contractors Association (“UCA”) had 28 active members at the time suit was filed in 1972, and between 1972-75 lost half of that membership. The remaining defendant association, Pennsylvania Excavating Contractors Association *342(“PECA”), dissolved in 1972; the record is silent as to its membership. Notably, both UCA and PECA have consistently delegated their negotiating responsibility to CAEP.
As a result of contracts entered into pursuant to negotiations with CAEP and GBCA in 1961 and subsequent negotiations with contractor associations, all participating contractors have passed on the direct managerial control of their application process to the hiring hall operated by Local 542. The contractors did not, however, relinquish power to affect the union’s operation of the hiring hall, for the contract terms remained fully capable of enforcement in the event they were violated by those operating the hiring hall and the contract instituted a grievance procedure by which an appellate tribunal consisting of an “Employer Representative, a Union Representative and an Impartial Chairman appointed jointly by the Employer and Union .” would decide whether a complaining job applicant was aggrieved “with respect to the functioning of [the] hiring agreement.” Article II, § 2(m). Although a vast majority of the employers are not and have not been active members of the defendant associations, the negotiations conducted by those bodies have established a standard to which the unaffiliated contractors may conform.
The hiring hall system is on its face neutral and purports to create a bona fide seniority system. Plaintiffs’ allegations of discrimination are not directed against the hiring hall system per se but against the union’s alleged intentional refusal to follow their own hiring hall rules, thus causing intentional discrimination against and a discriminatory adverse impact on minorities. Plaintiffs also allege discrimination particularly in admission into the union and admission to the job referral process. This allegation, although not attacking the hiring hall procedures set forth in the contract, does attack practices which would render the use of the hiring hall intentionally and otherwise discriminatory and contradict its bona fide seniority appearance. The foundation of this case rests on the statistical, documentary and testimonial evidence of discriminatory departures from and applications of the union hiring hall system.
B. The Philadelphia Plan
On September 24, 1965, shortly after Title VII went into effect, President Lyndon Johnson issued Executive Order 11,246 10 prohibiting discrimination by contractors with federal contracts in excess of $10,000 and requiring affirmative action to ensure non-discrimination. The Department of Labor and the Office of Federal Contract Compliance (“OFCC”) have principal responsibility for the enforcement of this Order.
Since much major construction work is done with at least partial federal funding, the federal government expressed its concern that the Order be complied with in the construction industry in the Philadelphia area. Toward the end of 1966 Bennett Stalvey, Director of the Regional Philadelphia Office of the OFCC, met with officials of Local 542 and later with officials of the CAEP and GBCA, advising them that it was the function of his office to see that the Order was complied with.
At this point, there was a clear disagreement between the federal government (OFCC), which wanted a higher percentage of minority individuals on construction jobs, and the construction industry, which wanted to use its traditional methods of employment. In the Philadelphia area (Philadelphia, Bucks, Delaware, Chester and Montgomery Counties) this matter took on added significance. Bennett Stalvey, in the fall of 1967, developed what was known as the “Philadelphia (Pre-Award) Plan,” under which the requirements of the Executive Order would be implemented. A low bidder on a construction project involving certain levels of federal funding would be obliged by the OFCC to maintain a specific level of minority representation in his construction *343workforce.11 The contractors, however, refused to make the specific commitments and hence in the spring of 1968 federal funds were withheld. Although Stalvey testified that there were several instances in which funds were withheld, the record specifically reveals only one low-bidder contractor-employer who was to use operating engineers out of 542’s hiring hall and from whom funds were withheld — Kiewit Sons Company of Omaha, Nebraska, a company which was to undertake a major highway construction project in Pennsylvania. Kiewit was not at that time a member of any regional contractor association, although it had been a member of CAEP by the end of 1967. Stalvey testified, and I find, that a total of at least $30 million in funds was withheld from highway construction programs through the spring of 1968.
A program was then developed which would accommodate the union in serving as an exclusive hiring hall and at the same time assure a substantial minority participation in the operating engineer trade. After a June, 1968, meeting in Washington, D. C., among officials of 542, their international representatives and OFCC officials, an “Affirmative Action Program” was executed on July 17, 1968. This program, as explained by Howard Minckler, a CAEP official, specifically included language urged by the contractor associations expressly making it in lieu of any other affirmative action plan, i. e., the Philadelphia Plan. Local 542, UCA, CAEP, and PECA were the original parties to this agreement, although by August, 1968, GBCA became an additional party. On the first page the agreement states:
The parties to this agreement have made a detailed analysis of employment of minority group workers in the classifications covered by this agreement and have determined that, in the area governed by this agreement, approximately 650 members of Local 542 are minority group members out of a total membership of 5400. [Emphasis added.]
The affirmative action thrust of this program was to establish an apprenticeship program into which minority members could be recruited and trained. In addition, there was to be a program for retraining journeymen, particularly minority journeymen. The federal government accordingly approved the program in lieu of the Philadelphia Plan. Thus federal monies were released and no money has since been withheld. The statement that 650 (12%) of 5400 members of Local 542 were minority members was viewed by defendants to be an important factor in obtaining federal approval of the substitute Affirmative Act Program and thereby eliminating compliance with more stringent federal minority manpower requirements. The “detailed analysis of employment of minority group workers” by Local 542, CAEP, UCA, PECA and eventually GBCA was, however, grossly exaggerated and totally in error.
In order to appreciate the significance of the overstatement of minority members in 542, it seems useful to outline the history of such inaccuracies. In the early 1960s, Robert Walsh, Business Manager of Local 542, told the Philadelphia Human Relations Commission that by his guess the figure was 500 of 5000. In 1966 Local 542 officials Walsh, Cahill and Ciavaglia informed Bennett Stalvey that the number was 800-900 of 5000. Later Minckler in a meeting with Stalvey stated that the number was “very large.” Based on the figures stated to Stalvey, the October 27, 1967 copy of the revised Operational (Philadelphia) Plan lists the minority composition as stated by the union at a possible 800 — 900 of 5000. In a 1967 report to the EEOC the union estimated a figure of 650 of 5000 stating, inaccurately, that most of its members “including Negroes and Spanish Americans have permanent employment” and therefore do not use the referral system. In 1968, Robert Emriek, formerly an official in the union but at the time coordinator of the Registered Apprenticeship program, told EEOC *344investigator James Nunes that his estimate was 600 blacks in District I and 50 more in other districts. In that same year the union wrote to Robert Bartlett, Pennsylvania Secretary of Highways, that 11 — 12% of the union membership were minority group members. This served in part as the basis for Bartlett’s objection to the withholding of funds in 1968. See note 11 supra.
Finally, by 1969, the stated level of minority representation began to recede into reality. Stalvey was told in a February 18, 1969 meeting with union officials Robert Walsh, Homer Dawson, Joseph O’Donoghue and their attorney, Martin Vigderman, that there were 400 minority members among a total of 6000 members.12 By December of 1970, the union had ascertained that the number of minority members as of January 1, 1970 was 259 of a total membership of 6128. This number was reported to the EEOC by a document dated December 31, 1970, and signed by Robert Walsh and was referred to in correspondence to the OFCC by Homer Dawson, local union president, in December of 1971.13
It is not acceptable to describe the repeated gross inaccuracies as merely incorrect guesses. Those defendants who signed the Affirmative Action Program agreement stated that the figure was arrived at after a “detailed analysis.” It obviously was not. Furthermore, Mr. Ciavaglia, the union official in charge of the hiring hall, with the assistance of Mary Kelly, a union employee, undertook a study in 1969 or 1970 based on records and their own knowledge of the membership, and concluded that there were an estimated 200 blacks in District I of Local 542. The ultimate unionwide count of 259 minority members as of 1970 was achieved by a similar method, one which was available throughout the period during which the inaccuracy was perpetuated, i. e., through “personal knowledge and identification by the business agents and officials of the Union and by contacting the last known employer when not known by the agents.” In addition to these distortions, 542 failed to include in their semi-annual reports to the OFCC information on where and in what capacity minority operating engineers were employed although such data was required by the Affirmative Action Program Agreement.
While it is conceivable that in one instance the union could have inadvertently made a significant error in overestimating the number of minorities in the union, it is incredible that errors of this magnitude could have occurred consistently by any mere coincidence. I find, and the record permits no other plausible inference, that the repeated overestimation of the number of minority individuals in the union, the failure to file complete semi-annual OFCC reports with the prerequisite data, and the other related activities of the union on this issue of the “Affirmative Action Program” were part of a deliberate scheme of Local 542 to deny to the federal government accurate information on the percentage of minority individuals in 542. This scheme was part of an effort to have federal funds inappropriately released while at the same time permitting 542 to keep the minority representation in the union at a far lesser level then it otherwise would have been if the Philadelphia Plan had been applied. Only a finding of discriminatory intent can explain this subterfuge. The Philadelphia Plan would have resulted in a higher number of minority group members in the union and in the related industry. I find further that the Philadelphia Plan would have been adopted by the defendants if the fraud, deception and scheme of Local 542 had not been pursued by the willful presentation of substantial overestimates of the number of minorities in the union.
*345Any argument that, because the union alone had primary access to the membership data, the contracting associations CAEP, UCA, PECA, and GBCA were not at least reckless participants in this scheme, I find to be devoid of merit and patently incredible. When these same contractors sought the release of the 30 million dollars they showed no hesitation in signing the statement that they “made a detailed analysis of minority group workers . . . .” At the time of this certification, there was no real suggestion that any of the signatory contractor associations had any doubt about the accuracy of the data or had merely limited knowledge of it; instead, they endorsed it. In this respect their posture is like that of an accounting firm which has certified that it has counted petty cash monies or has verified bank statement deposit records when, in actuality, it has done neither. In the instant situation, the record reveals a total absence of concern on their part regarding the accuracy of their representation to the federal government in seeking the immediate release of at least 30 million dollars and doubtless the release of millions more to follow. The prospect of deriving such an immediate and substantial financial benefit from the federal coffers allowed them to become willing parties to the scheme by capriciously certifying “facts” in anticipation of the government’s reliance on them: Having sought to enrich their members with substantial profits, it is now too late to cry innocence and cast the blame elsewhere. These were no innocent prognosticators who were misled by the union’s scheme to give inaccurate information. Under these circumstances, I find that these signatory defendant associations are consequently estopped from repudiating their certification after their members have had the opportunity to compete for and after many have received the released funds.
C. The Benjamin Franklin Programs I and II
In 1968 James Longacre, Executive Director of the Pennsylvania State Council of Operating Engineers (a private organization), formulated a six-month training program consisting of on-the-job instruction and classroom instruction for 100 hard core unemployed males from eastern and western Pennsylvania. While there were both white and black trainees, the majority of the members were minority individuals. An agreement was reached between the Council and the Pennsylvania Department of Labor to implement this plan. Local 542 had some measure of responsibility for final selection of the participants from eastern Pennsylvania even though (1) it was not asked to contribute to the program’s funding, which came from the federal government ($231,135) and the Commonwealth of Pennsylvania ($542,884.27), and (2) the Council of Operating Engineers and the Pennsylvania Bureau of Employment Security were fundamentally responsible for recruitment. Some equipment was donated by contractors, including Glasgow, Inc., but most was paid for in rentals from the program’s funding. None of the associations or contractors were parties to the agreement or responsible for the Benjamin Franklin I (BFI) endeavor and none received federal funds for the administration of the program.
The recruitment and selection activities by 542 were conducted through the JATC and particularly by George “Al” Holland, a black operating engineer who was a union business agent. Two minority applicants, Ronald Richardson and Howard Williams testified that they were told by Holland at the recruitment stage that upon successful completion of the BF program they would obtain steady work. This was repeated during training and at graduation exercises along with the assurances of union membership and journeyman’s pay. Even a graduation handout stated that graduates would receive highway construction jobs.
Of the 122 original trainees, 81 graduated from BF I. Fifty eight of these were black, 23 white; of this group 37 blacks and 7 whites were from eastern Pennsylvania. None of these graduates was • given any credit for hours spent in training. This is a matter of no small significance since the *346number of hours of experience, described earlier, is supposed to enhance placement in the hiring hall system. These graduates were placed on the Group III (lowest priority) out of work list. Although the director of the Benjamin Franklin Program, Mr. Nygard, testified that the graduates needed “follow through” (on-the-job assistance), none was provided. Mr. Walsh testified that it was his expectation that the graduates would work in yards, mills, or trucking companies but not in construction. As will be detailed below, many graduates received little work opportunity.
Toward the close of 1969, a Benjamin Franklin II (BF II) program was agreed upon, again at the instance of Mr. Longacre and with the approval of the Pennsylvania Department of Labor. The same format was used for recruitment and selection except that a tenth grade education qualification and screening tests were required of entrants and the Pennsylvania highway contractors were themselves required to contribute $80,000 statewide in cash or equipment as a condition to continued eligibility for bidding. The contractors formed Construction Training, Inc. in order to make the financial arrangements. Each contractor was to contribute to Construction Training, Inc. a base amount of $200 plus a rate of $.04 for each hour of time worked by an operating engineer or apprentice between March and October 30, 1970.
As with the BF I program the defendants had no part in formulation of BF II. George Holland did, however, have major responsibility in the process by which the union recruited applicants. Holland made statements to John Henry Owes that he would receive 2500 hours credit, a parent body book, and future on-the-job assistance if he graduated. Holland also made similar statements to Thomas Taylor, particularly that he would become a journeyman and earn $300 to $400 per week.14
The training for BF II took place at the same location as the BF I training, Resica Falls, Monroe Gounty, Pennsylvania. Of the 65 selected, 56 were graduated; of these, 46 were black, 24 from eastern Pennsylvania. Once again none of the graduates were given time credit, no one received a journeyman’s status, and after graduation only a few minority graduates accumulated a significant number of hours. In BF II, as in BF I, the associations, contractors and unions did not receive any funding.
The immediate net result of the Benjamin Franklin Programs was to increase the number and percentage of minority registrants in Group III status. I find that many of the BF trainees had been led by the union to believe they would be put in Group I rather than in Group III status and thus have the advantage of better job opportunity. In the years to follow that training, many though not all would fail to achieve the hours necessary to become members of the union or to enter Group II or Group I. The Benjamin Franklin Programs figure heavily in the present action because all of the named plaintiffs are BF graduates.15
*347D. Procedural History16
Having outlined briefly the hiring hall system of 542 and having discussed some of the important circumstances underlying this suit, particularly the Benjamin Franklin training programs and the participation of defendants Local 542 and contractor associations in a program substituting for the Philadelphia Plan, attention may meaningfully be directed to the procedural history of this case.
The first major legal challenges to the alleged discrimination began with two charges filed before the EEOC, one by an EEOC Commissioner Vincent T. Ximines on July 29,1968, and one by three BF I graduates, Raymond Williams, Willie McKay and Donald Muchison, on June 17, 1969. EEOC investigator James Nunes investigated both complaints. Service of Ximines’ charge upon Local 542 was made on October 3, 1968 and service of Williams’ charge was received on November 28, 1969. Ximines’ charge was directed to the breadth of 542’s procedures:
(a) Respondent discriminates against Negroes in referrals for jobs.
(b) Respondent discriminates against Negroes by limiting and classifying its membership in a manner which deprives Negroes of employment opportunities.
(c) Respondent discriminantly fails or refuses to admit Negroes to apprenticeship programs.
(d) By these and other acts, Respondent has discriminated and continues to discriminate against Negroes because of their race in violation of the provisions of Sections 703(c) and (d) of the Act.
The pro se charges by Williams, McKay and Muchison are identical to one another. They allege discrimination in the union’s referral practices:
I was recruited into the training program by a representative of the International Union of Operating Engineers, Local No. 542 and upon completion of the training program I was promised employment by the Union. The training program under the sponsorship of the Union ran for a period of six (6) months. During this time I was trained to operate the following heavy equipment: Bulldozers, Graders, Front End Loaders, Back Hoes, Post Hole Drills, Core Drills, Scrapers, etc. I was graduated as a qualified operator on the above equipment. Since my graduation two weeks ago I have been to the Union Hall every day to be referred, with no [undecipherable word] results. I feel that I am being discriminated against because of my race (Negro) and for no other reason.
On August 30, 1971, the Philadelphia District Director of the EEOC, Ralph A. Allen, advised Local 542 in separate letters of the decisions of the EEOC that reasonable cause existed to believe (1) that Ximines’ charge was true and (2) that the charges of Williams, McKay, and Muchison were true. The three claimants, in the decision in their case, were deemed entitled to receive the EEOC’s decision dated August 6, 1971, on the Ximines charge. The letter of decision on their own case stated that, as members of the class harmed by the discrimination alleged by Ximines, the three individuals were entitled to such notice:
The Companion case . . . concerns a charge lodged against Respondent by an EEOC Commissioner. Pursuant to our Procedural Regulations, 29 C.F.R. 1601.25b (June 18, 1970), 35 FR 10005, Charging Parties will be entitled to receive copies of our decision in [the companion case] both because they are currently aggrieved by practices found to be unlawful in that decision, and because they were members of the class of potential Negro members which was aggrieved by several of Respondent’s unlawful employment practices at the time the Commissioner’s charge was filed.
*348Right to sue letters dated October 14, 1971, were sent to Williams and McKay.17 Within 30 days thereafter, on November 8, they and the other named plaintiffs in this suit filed their class action complaint with the district court. On January 3, 1972, a class of plaintiffs and defendants was certified, as described above.
During the. discovery stage which followed there was an unexpected and very unfortunate development in retaliation for the filing of plaintiffs’ suit. On June 19, 1972, Marion Eaddy and John Dent, both named plaintiffs, were attacked by white operating engineers in the union hiring hall offices. Dent’s testimony, which I credited in my previous opinion at 347 F.Supp. 268, 275-76 (E.D.Pa.1972), aff’d, No. 72-1901 (3d Cir., May 21, 1973), was particularly graphic. Marion Eaddy had been hit over the head from behind with a chair. Two men then continued punching and kicking him as he lay on the floor. Dent pulled one man away from Eaddy and shouted for help from the others in the hiring hall who included the union’s business agent, Mr. Ciavaglia, his assistant, a C branch agent and twelve to fifteen u,nion members. Dent himself was then struck over the head with a chair by a white operating engineer. The police were never called. The victims were not forewarned or aided by those witnessing the occurrence. Although Ciavaglia testified that he did not know the attackers, one of them had been in his office just a few minutes before.
On the following day the violence increased. Four or five pickup trucks in each of which were several white men, some of whom Dent recognized as operating engineers, came to the hiring hall en masse shortly after working hours. Despite the presence of a policeman, Dent, Eaddy and Cleveland Allen were beaten by groups of white operating engineers in front of the hiring hall. 347 F.Supp. at 276-77.
After an extensive examination of the facts and law this court entered an order enjoining pendente lite Local 542 and all those acting in concert with or on behalf of Local 542 from interfering in any way with or retaliating for plaintiffs’ exercise of their federal statutory and constitutional rights to institute their employment discrimination action.
Following four years of discovery, I entered an order bifurcating the trial into two stages: the first on the issue of liability and the second, if liability was found, on the issues pertinent to recovery of damages.
E. The Delays and Contempt Issues Involving Abraham Freedman, Esquire
During the course of the trial which began on January 19, 1976 and continued through June 28, 1977, there were two other unusual occurrences deserving mention. Local 542’s counsel, Abraham Freedman, Esquire, was twice cited for criminal contempt for his courtroom conduct. 73 F.R.D. 551 (E.D.Pa.1976), aff’d, 552 F.2d 498 (3d Cir.) (affirming both criminal contempt orders), cert, denied, 434 U.S. 822, 98 S.Ct. 67, 54 L.Ed.2d 79 (1977); 73 F.R.D. 544 (E.D. Pa.1976) (denial of motion for stay of action pending Freedman’s appeal of first criminal contempt order). The first contempt order followed a lengthy and undirected reading by Mr. Freedman of Bennett Stalvey’s deposition during his cross-examination. After four transcript pages of such unfocused reading, plaintiff’s objection was sustained. Mr. Freedman responded by objecting to the court’s ruling. He stated his intention further to state his ground for objection. He was ordered not to do so with the assurance that he would be the beneficiary of all possible grounds. After numerous repetitions of the order, 73 F.R.D. at 548-50 (Appendix A), Mr. Freedman, wilfully persistent in his course, was held to be in contempt. Freedman received a thirty-day sentence which has since been modified to a $500 fine upon Freedman’s motion for re-sentencing.
The second contempt order was a result of Mr. Freedman’s calculated dilatory tactic to continue reading from Samuel Long’s work record (in evidence as an exhibit) de*349spite an order to cease such reading. 73 F.R.D. 551 (E.D.Pa.1976). For this refusal Mr. Freedman was fined in the amount of $500.
From the very beginning of this lawsuit Mr. Freedman’s tactics of delay and reprehensible conduct were wilfully designed and executed with the intent to obstruct and preclude a proper judicial determination of the real facts of the case. It is my judgment that he sought to subvert this intensely controversial matter to a mistrial so that years of investigation and discovery would have to begin again. Although his conduct was as reprehensible and as irresponsible as any I have seen during thirteen years as a trial judge, nevertheless, with patience and hopefully discernment, I have tried to disassociate those tactics from the essential fact finding process here.
Fortunately, for all the litigants including his clients, Abraham Freedman was not successful in aborting the trial process; however, he did succeed in making most difficult the presentation of those issues which would have been easy to discern from a fact-finding standpoint and which could have been ruled on and adjudicated promptly. As to the more complex matters in the case, he persistently tried to make it impossible to comprehend their nuances and to preclude any rational presentation of the ease. By reason of this daily persistent level of counsel irresponsibility, discovery, trial and adjudication have extended far beyond any reasonable time span which any case would warrant — even one involving a most intense vigorously litigated controversy-
Perhaps the tone of the trial can be best conveyed by the opinion of the Court of Appeals on the two contempt citations. Commonwealth of Pennsylvania v. Local Union 542, 552 F.2d 498 (3d Cir.), cert. denied, 434 U.S. 822, 98 S.Ct. 67, 54 L.Ed.2d 79 (1977). After evaluating Mr. Freedman’s conduct, the Court of Appeals reviewed cases where similar conduct of non-lawyers was held to be inexcusable and then the Court noted:
If non-lawyer Seale’s defiance of a judge’s order was inexcusable, how much more so should be the conduct of appellant Freedman, a seasoned trial lawyer. We agree with the Seventh Circuit, and hold that a trial attorney’s belief that certain action is necessary to protect the record for appellate review does not excuse his deliberate defiance of the trial judge’s explicit and repeated orders. The phrase “preserving the record for appeal” is not a talisman that absolves a lawyer from his usual obligation to comply with a trial judge’s direct orders.
* * * * * *
Freedman’s conduct was an affront to the trial judge’s authority to control the proceedings in his own courtroom. An attorney who, in deliberate disregard of seven direct and explicit orders by the trial judge, pursues a course that he determines to be in the best interests of his client, offends the dignity and authority of the court and thereby obstructs the administration of justice. United States v. Seale, 461 F.2d 345, 371 (7th Cir. 1972). To hold otherwise would be to strip trial judges of their power to supervise the proceedings before them, and to clothe counsel with the authority to conduct trials in whatever manner they deem appropriate. Furthermore, we note that Freedman’s conduct was obstructive because it resulted in a wholly unnecessary and not insignificant delay of the trial. For this reason, too, we conclude that his conduct arose to the required level of disruptiveness.
sf: * * * * *
Freedman’s action was intentional, constituted misbehavior which caused an actual obstruction of the administration of justice, and occurred within the presence of the court.
* * * . * *
Freedman concedes that his conduct did not vilify the judge, and we cannot agree that the judge’s comments demonstrate a bias against Freedman. The judge’s description of appellant’s conduct was nei*350ther acid nor sarcastic; it was simply accurate. The trial judge exhibited patience and restraint, and did his utmost to preserve order and decorum; he did not engage in wrangling or bickering, and used the summary contempt power only as a last resort . . . . Appellant [Abraham Freedman] mistakes judicial disapproval for personal pique. A judge who objectively expresses his antipathy toward contumacious conduct does not thereby disqualify himself from adjudicating the contempt under Rule 42(a). Under the circumstances of this case, we see no need for another proceeding to adjudicate what one judge has already witnessed. Summary disposition, we hold, was appropriate.
After the contempt citations had been sustained and on the last day for modification of his thirty day sentence, he filed a petition requesting that he not be sent to jail — even though the sentence had been affirmed and no further judicial relief seemed available. I modified the prison sentence by vacating the imprisonment and imposing a fine of only $500. This was done not because there were any fundamental mitigating factors but solely because if, at his present age of more than three score and ten, he has never learned or has now lost totally the recognition of the standards of permissible conduct for lawyers, then 30 days in jail would not be a true deterrent for the future. Thus the jail sentence was not eliminated because of any view that it was ever unfair, for I am still confident that Mr. Freedman’s conduct was as inexcusable as it was so often disgraceful.
F. The Named Plaintiffs as of Certification
As has been noted all twelve of the named plaintiffs are graduates of either Benjamin Franklin I or II. Because of the importance of the characteristics of these named plaintiffs in determining whether they satisfy the class action requirements and rule 23(a), the plaintiffs will be described with reference to some basic features.
At the time of the certification, on March 13, 1972, one of the named plaintiffs (Marion Eaddy) was a listee in Group I, five were in Group II (Randolph Hughes, William Bostic, Ronald Richardson, Ronald Crawford, Willie Frank Gilchrist) and the remaining six in Group III (Raymond Williams, William McKay, Arel Brownlee, Kenneth Howard, Alpha Christmas, Clarence Winder).
One, Marion Eaddy, a graduate of BF I, did become a member of Local 542’s parent body. Two of the named plaintiffs, Randolph Hughes and Alpha Christmas, also BF I graduates, were applicants to the Joint Apprenticeship Training Program, although neither had received a final disposition of his application. The JATC files of Hughes and Christmas indicate no disposition but simply make reference to the Benjamin Franklin program. All plaintiffs were alleged to have been victims of the referral practices and practices relating to admission to Local 542 and discriminatory advancement in their profession.
The hours worked by named plaintiffs range from none at all to over 2500 in the period between their BF graduation and class certification. Three, Raymond Williams, William McKay and Clarence Winder never received registrant books, the first and lowest step in attaining entry to the union.
Although it is clear that the Group III plaintiffs are different from the Group II and Group I plaintiffs in terms of their supposed priority in referrals, I find there is no conflict between such plaintiffs for purposes of trying the legal issues of this suit. Plaintiffs are not attacking the hiring hall system as constituted but rather are attacking practices which, while forming no part of the hiring hall system’s stated foundation, nevertheless have a profound impact on its operation.
G. Statistical Evidence
An important part of plaintiffs’ case lies in its proof of discrimination in membership in Local 542 and its proof of discrimination in hours and wages of minority union mem*351bers. Plaintiffs’ proof on both issues is based in part on a statistical probability analysis. See International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977). For the sake of orderliness we will take the statistical proof of membership discrimination first.
Plaintiffs’ expert, Dr. Bernard Siskin, is an Associate Professor at Temple University and Chairman of the Department of Statistics. He has written numerous articles and a text book in the field of statistics and has particular expertise in social statistics. At trial he presented a number of studies of 542’s composition, ascertaining, inter alia, the number and proportion of minority union members, the pool of minority persons available for operating engineer work, the number and proportion of minority entries from 1966 to 1975, the numbers of minority persons entering the union via the various available means, and the comparative hours and wages of minority members. From such studies he was able to draw conclusions relating to the likelihood of discrimination in the 542 hiring hall system. This case, like most cases involving statistical proof, involves among the expert witnesses some evidentiary conflicts which require the fact finder in turn to make findings of credibility. Upon balance I find Dr. Siskin’s testimony to be credible, persuasive and accurate on these subjects and on the other subjects of his testimony. The testimony of the other experts who differ does not cause me to repudiate or modify my finding as to Dr. Siskin’s credibility.
1. Membership Disparities
Based on the computer tapes of a “Master List of Active Members” provided to Dr. Siskin by the union (with key punch errors being corrected through the annual pension files), Dr. Siskin determined the union membership to be 6,051 as of December 31, 1971. This figure includes the parent body, branches A through D, and the Registered Apprenticeship enrollees. (Holders of registrant books are not counted in this tabulation and are not considered by the union as members.) There were two hundred and thirty-five black members. Thirty-three more were members of another minority. Thus the composition among active members of Local 542 at the end of 1971 was 3.9% black. When all minorities are considered, the minority composition of the union was 4.4%. These 1971 percentages are very nearly identical when pension welfare, and honorary members (inactives) are included in the membership definition. The following table reflects the composition of 542, including pension, welfare and honorary members, for the years 1966 — 71:
Based on the 1970 census data for the area covered by Local 542’s jurisdiction, the total population is 7,729,115 of which 888,-370 or 11.5% are black and 33,073 or .4% are members of another minority group. Limiting the population to males between the ages of 18 to 65, in order to define more precisely the pool of potential applicants, the figures become 11.0% black and 11.5% minority. The 11.0% and 11.5% figures are a conservative statement of the available black or minority labor pool. This was Dr. Siskin’s conclusion and I agree. Siskin’s best estimate, however, was 12.7 to 13.4% black and a total of 13.2 to 13.9% minority, taking into account by his calculations labor force participation rates, the census under-count, education and occupation and disregarding particularly the suggested factors of automobile and telephone ownership.
Using the conservative 11.0% and 11.5% result, the membership of Local 542 as of 1971 was grossly disproportionate to the 3.9% black and 4.4% minority percentage in the labor pool. The likelihood that such a disparity would occur by chance is less than *3521 in 100 trillion, less than 10 "23. Of course, if Siskin’s greater percentage figures, which I find on a preponderance of the evidence to be accurate, were used, the disparity would increase still further. Notably the disparity remained extremely significant as against the conservative labor pool figures as of 1975 when the minority percentage (four years after the initiation of this suit) had risen to only 4.8%.
2. Discrimination in Entry
In addition to demonstrating this gross disparity, plaintiffs’ expert, Dr. Siskin, compiled data principally from union exhibits indicating the ratios of minority entry into the union (parent, A, B, C, and D branches and the RA program). During the period 1966-1971, 2601 new members entered 542. Minorities entered at the following rates:
Blacks Minorities
1966 5.1% 5.5%
1967 2.6% 2.8%
1968 5.8% 5.8%
1969 8.2% 9.8%
1970 3.8% 4.8%
1971 6.8% 8.7%
The total black entries for the union as a whole were 143 or 5.5% over the course of these years and the total minority number was 166 (6.4%). The direction of the entry rate (increasing or decreasing) fluctuated during this period, so it is not possible to deduce with absolute certainty any definite trend, although in the last three years before suit (1969 — 1971) the average entry rate was 6.4% black and 7.9% minority. The difference between this average and the conservative labor pool percentage is statistically significant at less than 1 in 1,000,000 (i. e., the probability that the difference can be explained by chance is less than one out of one million). On the basis of random entry one might reasonably expect almost 300 minority entries between 1966 and 1971. The difference between this expected number and the lesser actual number is 133— 44%. By the end of 1974, the minority membership of Local 542 was 4.3% black and 4.8% minority out of 6725 members. As will be developed below, there are no other factors which would fairly require a finding that this gross disparity is not the result of discrimination. There is no sufficient evidence to explain that this disparity occurred because of any valid job-related qualifications not possessed either by members of the minority labor pool or by actual applicants for union membership.
Even after suit, from 1972-1974, the minority entry ratio into the union is at 7.4%, not far from the 6.4% pre-suit figure applicable between 1966 and 1971 and less than the 7.9% rate between 1969 and 1971. In 1972, 5.9% of entrants were minority; in 1973, 8.1%; in 1974, 8.0%; and in 1975, based on union data after excluding reinstatement and withdrawals, only 1.5% (adjusted from 2.8%).18
In the data given above, C and D branch members were properly included in assessing the extent of 542’s discrimination in part because the C and D members are indeed union members and because transfers can be affected from C and D into the parent body construction trade, thus affecting the general measure of availability of entry into the construction trade and the employment prospects of the plaintiff class. (The entry data for the years 1972 — 1975 does not include D branch members.19) •Plaintiffs’ arguments that Local 542 has discriminated are indeed based in part on the assertion that the union maintained primarily white entry methods and primarily minority entry methods. For purposes of analyzing this contention the data on entries into the union will be broken down (in rounded figures) into branch or division.
*353
The overall average entry rate from 1966-71 is:
Union A&B&P C&D RA
Black • 5.5 6.4 4.6 4.3
Min. 6.4 7.0 6.0 5.0
The overall average entry rate from 1969 to 1971 is:
Union A, B, P C, D RA
Black 6.4 8.3 3.3 6.4
Min. 7.9 9.4 5.8 7.5
For each of the three year averages (1969-71) the disparity from the minority labor pool percentage is statistically significant (less than a 5% probability of chance occurrence). For the union as a whole, including C and D branches, during this three-year period the disparity is significant such that the likelihood of these overall figures occurring by chance is less than 1 in 1,000,000.
In the period from 1972 to 1974, 7.4% of entries into 542 were minority members: 8.0% parent body, A, or B; 1.7% C Branch; 18.2% RA.20 The very high RA figure comprises entrants into the four-year program. If instead of examining the number of minority entrants into the RA program we look to the number of minority graduates moving from RA status into the operating branch for this period, the 18.2% figure diminishes to about 5.9%. This latter percentage is in many respects the most meaningful in measuring the minority participation of registered apprentices. Furthermore, as the minority RA entrants have increased in this period, the minority C Branch entrants have diminished somewhat, so that apparent increases in participation by minorities in the RA program are not absolute increases. If the RA minority graduate figures are used instead of RA entry figures, the overall percentage of minority entrants into Local 542 during 1972-74 is 4.9% instead of 7.5%; this makes a decrease since the filing of this suit in 1971 when the average rate of minority entry between 1971 and 1966 was 6.4%. Although the RA program will be discussed separately under the heading “JATC,” it is clear that a careful evaluation of the statistical proof reveals continuing discrimination. Thus as of 1976, two years after A and B branches were abolished, the Parent Body was 5.8% minority, the C Branch 3.9%, and the Registered Apprentice Program (including entrants) 12.7%. Excluding inactives the percentage of minority union members (counting RA entrants) at the outset of 1976 was 4.87%. The fluctuations in methods of entry strongly suggest manipulation. This is corroborated by other proof.
3. Disparities in Hours and Wages
The second major component of plaintiffs’ statistical proof concerns the proportionality vel non between the hours worked and wages received by white members and by those minorities who were able to become members of Local 542. Dr. Siskin performed two basic studies inquiring into this matter. In the first he compiled from the union’s Master List of Active Members all the hours worked by whites and minorities during 1969-71 and estimated by a standard formula of hours worked in recent years just who was in Group I, II or III. In the second, using the Master List he examined the hours worked and wages obtained during 1972, making correlations to the out-of-work list and the respective operating branches of 542 so as to compare even more closely like groups of white and minority workers. Both studies excluded consideration of C Branch and D Branch and registrants.
The result of Siskin’s first study based upon health, welfare and pension records is indicated in the following table.
*354
When one compares the minority labor pool percentage (11.5%) to the minority member percentage of hours worked, the disparity is still greater than that revealed by the comparison between the minority labor pool and the 542 operating membership.
Plaintiffs’ first study sought to account for differences in age, district, branch, seniority, and out-of-work list status by use of a regression analysis. After accounting for these factors, Dr. Siskin nevertheless concluded that white operating engineers in the studied branches worked 109 hours per year longer and at higher rates than minority members on account of race. Siskin concluded that the racial disparity was statistically significant and would not be decreased by the addition of other variables. This means that minority members received $717 per year less than whites. In addition to this loss, Siskin concluded that minorities lost on the average $262 per year due to maldistribution among branches.
Upon receiving master list data for 1972, including for the first time exact wage information and group status, Dr. Siskin repeated his analysis for that year, excluding from consideration all those individuals on the list who began during the course of 1972. Siskin also analyzed in this study the raw wage and hour differential among the branches and among out-of-work listings. Once again, only Parent, A, B, and RA branches union wide were analyzed:
The average racial difference in wages by branch and group list unionwide is correspondingly as follows: 21
The average overall disparity unionwide in 1972 hours and wages is revealed by the following table:
*355After also subjecting the raw 1972 results to a regression analysis (so as to equalize factors of age, district, branch, seniority, and group out-of-work list status), Siskin concluded that whites on the average worked 97.5 hours more than minorities in 1972 and earned $749 more. This is indeed close to the 1969-71 estimated result. It must be emphasized that the disparity in average hours worked is not a result of the effects of low priority out-of-work list group or seniority. The discrepancies accrued while these factors were equal.
4. Referrals
Closely tied to plaintiffs’ proof of a differential in minority hours and wages is the proof of an arbitrary system of referrals. Siskin’s analysis on this subject was limited to District I, although in a separate analysis unionwide clustering of minorities with a limited number of employers was also shown to exist. First I will detail the District I study.
Siskin’s study of the District I referral system involved an analysis of seventeen out-of-work lists from among the various' groups (five from Group I, one from Group II, five from Group 1-A, and six oilers and RA lists). The seventeen lists were the remainder after eliminating all lists during that period with less than forty names. This was to assure statistical significance. Each list had been used for one of any of the months between 1969 and 1971. A rank was then assigned to each person on the list according to his position. Referrals were counted based on the first referral date marked in a listee’s work records. (A referral constituted any attempt to contact a worker including acceptances, refusals, or failures to achieve contact so long as noted in work records.) By computer, seventeen “selection” lists were created reflecting the actual order of referral.
By creating the selection list, Siskin was able to compare actual referral rankings to the work list ranking in order to determine the coefficient correlating the two lists positively, negatively or neutrally. The appropriate numerical correlation (the Spearman rank correlation coefficient (“r”)) ranges from “ -1” to “ +1”. A “ +1” correlation would mean that the two lists are identical; a “ -1” would mean that they bear a perfect reverse order correlation; a “0” correlation would mean that the relationship appears random. Based on Siskin’s analysis the r correlation coefficients for the seventeen lists are as follows:
Group I lists
List number r
# 4 .20
# 7 .55
# 10 .52
# 13 .46
# 16 .62
Group II
# 2 .08
Group I-A
#3 .22
# 5 .40
# 9 .37
# 12 .43
# 15 .54
RA & Oilers
#1 .24
# 6 .38
#8 .44
# 11 .44
# 14 .46
# 17 .45
While all of the lists except list # 2 from Group II were clearly on the positive correlation side, further analysis by Siskin revealed that virtually none of the lists reflecting actual referral rankings was significantly similar to the corresponding out-of-work list.
This conclusion was reached after calculating the “variance.” In the present context the correlation coefficient (r) when squared (r2) measures the variance in selection rank which is explained by the out-of-work list rank. The formula 1-r2 measures the variance in selection not explained by the out-of-work list. Of the seventeen lists examined, on only one was more than one-third of the selection rank predictable or explainable based on out-of-work list rank. On another list the position on the out-of-work list explained only .6% of the selection list ranking.
*356
The average for all lists indicates that 82.5% of variance is the result of factors other than order on the out-of-work list. Although there exists a possibility that selections based on skill could theoretically have created discrepancies in selection rank, plaintiffs point out, and I agree, that the low correlation on the single skill oilers and RA lists (a correlation very like that for the other lists) itself tends to disprove the theory that “skills” explains the variance.22
The next statistical test performed on the seventeen lists was to determine whether predictability of selections from out-of-work list ranking increased depending on the type of list. A table indicating the percent of explained variance within groups, with and without ranking those who were not referred at all, is below:
The percentage of explained variance is relatively small.
Siskin prepared yet another chart, an “expectancy chart,” indicating the probability of selection for persons listed in respective quintiles of each out-of-work list:
*357
As can readily be seen the probability of selection is not increased by being in the first or even second quintile on the out-of-work list, although someone in the fifth quintile of the out-of-work list is not at all likely to be selected among the first or second selection list quintiles. While this analysis in itself does not seek to identify race as the factor creating the lack of correlation, it confirms that the out-of-work list ranking is simply not the principal basis for selection. This corroborates plaintiffs’ claims of discrimination in the sense that it proves there is much room for arbitrary and standardless selections. When combined with the other statistical disparities considering the race factor directly, this correlation study aids the inference of discrimination.
Corroborative of the above data is the calculation of data indicating clustering of minority workers referred through the hiring hall system of Local 542. For the period 1969-1971, Dr. Siskin examined union pension records for all districts and determined that 69.6% of the 1488 employers reporting hours for the period reported no minority hours at all. The breakdown by years is as follows:
Twenty-three employers, according to Siskin, employed minorities during the three-year period as at least 10% of their total hiring hall employees. These companies were all in District I. They accounted for 32.41% of all minority hours in that district but only 5.96% of total District I hours. Further, these employers received 17.17% of District I minority referrals but only 7.43% of all District I referrals.
H. Other Proof
Proof of gross statistical disparities may itself constitute a prima facie case of intentional pattern and practice discrimination. International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977); Hazel-wood School District v. United States, 433 U.S. 299, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977). In this case there is more than statistical disparity. There is evidence of the local union’s intentional deceptions to avoid application of the Philadelphia Plan as well as evidence of deception in denying the Benjamin Franklin graduates union membership. These deceptions (particularly relating to the Philadelphia Plan) were, I find, engaged in with the purpose of appearing to advocate equal minority participation, and hence of assuring the flow of federal contract monies, while at the same time duplicitously resisting such participation at meaningful steps of implementation. Plaintiffs need offer no more than this to establish their class claim of intentional discrimination and bring the abstract statistics to life. Plaintiffs’ other proof consists, in part, of evidence showing specifically enumerated departures from customary entry *358methods and a relation between method of entry and race which cannot be viewed as coincidental. Plaintiffs also have presented testimony of twelve minority individuals relating their experiences in seeking access to the union hiring hall. Both types of evidence seek to prove a pattern of discriminatory conduct, or to corroborate other proof of discrimination, without relying directly on an expert’s opinion concerning the exact probabilities that the specific instances can be generalized. Such evidence has long been an important part of employment discrimination suits and assumes evidentiary significance by establishing or tending to confirm in the concrete that which statistics have abstractly assured was highly probable. Both types of non-statistical proof may create or corroborate an inference of discrimination.
1. Entry Discrimination
First for consideration is the proof of a variety of methods for entering Local 542 which have either departed from the customary entry avenues, and in the process have provided primarily white entry routes, or in the ordinary course of operation have resulted in the exclusion of a disproportionate number of minority applicants.
The initial formally permissible methods of, or steps toward, entry are through what are called A and B branch Organization, C and D branch Organization, C branch Miscellaneous, the Registered Apprenticeship Program, and the Registrant Program. As indicated earlier intra- and inter-union transfers may also be accomplished.
Entry through A and B Organization may occur when employers previously not committed to the union bargaining agreement decide to become committed. Their construction work forces are issued union books at that time without having to meet any qualificational criteria. (A and B branches were abolished as of December 31, 1974). Entry through C and D branch Organization is accomplished when non-union workers choose the union as their bargaining agent, usually by an election procedure. C branch Miscellaneous entry is the entry route for new workers hired by employers whose workforces had previously been organized. The Registered Apprenticeship Program, a four-year training program, provides an entry method for essentially unskilled would-be operators who meet certain qualifications and who pass written tests.23 Lastly, there is the registrant program, A and B. A-registrants were unskilled persons desiring to be operating engineers who simply registered with the union and obtained an A-registrant book. After achieving 2500 hours of experience membership in the union became available. This method of entry was eliminated in 1968. B-Registrants, prior to August 1, were those claiming to be experienced operating engineers who were issued B-registrant books. On August 1,1970, however, a written and field testing procedure went into effect to verify the qualifications of B-registrants. These tests have not been validated as to their job-relatedness.
Intra-union transfers render entry into the non-construction branches relevant to the issue of entry into the construction branches. Branch members may transfer into construction, by varying accounts, either after attaining 2500 hours, after attaining journeyman C branch status, or without restriction or specific rules. Transfers into the parent body could be effectuated from A branch after four years or B branch after three years. Inter-union transfers have also been permitted.
The following table, based on union records, indicates the minority composition of persons entering 542 through the above described methods between 1970 and 1975:
*359
From this data it is a simple matter to determine that most (60.6%) white entrants into 542, or into a category through which access to 542 could be attained, entered without being subjected to any standards of qualification. Only 16.4% of new minority entrants or affiliates, however, entered in such a way. Only 14.5% of whites entering or becoming affiliated with 542 entered through the Registered Apprentice Program although 35.8% of minority entrants or affiliates came into their positions through the RA program. A significant 47.9% of all new minority entrants or affiliates in this period became B registrants, while only 24.9% of whites achieved this status.
A closer breakdown reveals even more emphatically the minority underrepresentation in entries via B branch Organization:
Only seven out of 377 B Organization entries were minorities, a total of about 1.9%, much less than either the minority labor pool percentage or the percentage of minority union members.
Focusing on C-Branch also reveals diminutive minority participation in the new entries to that branch. Between 1970 and 1975 only 8 (1%) of 768 C Miscellaneous entries were minority members, and only 11 (4.5%) of 242 individuals entering by way of C Organization were minority members. By 1976 minority representation in C branch was only 3.9%.
As the minority rate of entry into and population of C branch has become relatively insubstantial, the minority representation among registrants has been fairly constantly greater than the minority labor pool percentage in the 542 geographical realm.
The rate of minority entrants into the registrant program between 1969 and 1975 was itself commensurate with labor pool percentages at 12.4% (103 out of 832). Of 44 A registrants entering Local 542 between 1970 and 1975 only 1 was a minority. Given these indicia, it is reasonable to conclude that there was among entrants a minority concentration in the B registrant non-member group.
*360 What is indicated by these discrepancies is that the union maintained channels of access which were in practice primarily white entry channels or primarily black entry channels. Those channels which were primarily black tended to be the ones with the strictest requirements. The B registrant entry and RA entry were the only methods requiring any testing.
The segregative channelling effect is explained somewhat by other non-legitimate methods through which entry or affiliations with 542 have been accomplished. Foremost among these methods is the simplest— direct and unexplained entry. Plaintiffs have assembled a list of forty-three white individuals who have entered 542’s parent body or A or B branches since 1970 without having been registrants and without having entered through organization, C branch, RA program or interunion transfer. No minorities have entered in this way.
Particular instances of direct entry also provide an insight into the union management’s role therein. The direct entry of one Nicholas D’Ambrosia, Jr., son of a former member of 542’s Executive Board, received emphatic treatment in plaintiffs’ proof. Testimony at trial reveals that Robert Walsh, business manager of 542, was informed that D’Ambrosia possessed no union book of any kind. Walsh instructed Ciavaglia to remove D’Ambrosia if this was found to be true. Ciavaglia determined that D’Ambrosia had been improperly referred as an oiler by one of the District I business agents, Joseph O’Donoghue. O’Donoghue refused to alter his position, while not denying its impropriety in his discussion with Ciavaglia. D’Ambrosia was never removed from his job. This case is illustrative of the numerous instances of nepotistic direct entries documented by plaintiffs. Although these instances cannot here be detailed, plaintiffs’ documentation as referred to in the Introduction is credited.
2. Individual Testimony
The experiences narrated below, while not necessarily presented to establish individual claims of intentional discrimination, reveal a pattern which is consistent with and therefore relevant to the pattern of resistance to union entry established by other evidence including statistical proof. It is with this characterization that the following redaction of individual testimony is provided. The order of presentation is primarily based on the date of union affiliation.
(a) Samuel Long
Years ago Samuel Long learned to operate heavy equipment in the Army in a segregated unit which trained at Fort Bragg. After service time on construction in Africa, Long was honorably discharged in 1944.
Long unsuccessfully attempted entry into 542 by seeking work three times between 1950 and 1956 from a master mechanic at the United States Steel plant in Morrisville. His father-in-law was an operating engineer there. (In the meantime he had been employed by non-union contractors at the Philadelphia Navy Yard until 1953 and later in private industry.) In 1956 Long was accepted into the union as a B branch member after finally obtaining employment at the U.S. Steel plant. In 1959 Long entered the parent body and in 1960 he was on the campaign committee for Robert Walsh, who, Long testified, campaigned for the position of Business Manager on a ticket opposing discrimination against blacks, Irish and Italians.
Long’s testimony does not reveal discrimination by the union against him. He apparently did not apply directly for union membership during his efforts to obtain employment at U.S. Steel. However, Long’s testimony as to the foundation of Mr. Walsh’s campaign for business manager provides some background evidence that the issue of racial discrimination was alive prior to the time period defined by the statute of limitations in this suit.
(b) Willis Fox
In April, 1958 Willis Fox obtained a job with a Philadelphia contractor, Amenio De Paul, and began work as a laborer. He was *361later assigned to baekhoes and other equipment before being laid off in June, 1961.
Fox’s initial attempt to join Local 542 was in August 1961. At that time James Grant, the hiring hall agent, informed him that the books were frozen — that too many men were out of work and his application would not be accepted. Fox returned to work for De Paul, and later worked for Mario Massi, first as a truck driver and then as an operator. In August 1964, Fox returned to the Hall again seeking entry into Local 542, and was told to return in October. In October Grant told him to return in November at which time he completed his application, paid $40.00 in fees, and was issued a B registrant book. Three years later, in 1967, he became a member of Local 542.
Fox testified that both as a registrant and B branch member his referrals were exclusively dirt work jobs operating loaders and pippins (relatively low-paying machinery); he received no referrals for cranes or graders. When he complained in early 1969 about this, Ciavaglia stated that upgraded referrals were impossible because so many men were ahead of him. Fox’s response was to find crane work through another local in New York, and over the next five years, spanning July 1969 through February 1974, he worked intermittently in New York as a crane operator. Fox’s association with Local 542 continued through that period.
Fox reported being sent out by Ciavaglia on a Heister roller to the Glasgow, Inc. on August 23, 1973. All three rollers at that job site were operated by blacks. The master mechanic had promised that Fox would be reassigned to higher paying equipment when it came in, but instead assigned whites over him. At that point Fox quit and returned to New York.
(c) Robert Ahmad
Robert Ahmad received training in sheet metal drafting before completing a correspondence course in surveying, maintenance, and the operation of front end loaders, dozers and scrapers from the Training Service Institute. His field training consisted of a one-month course in operating the front-end loader, dozers and scraper, and was completed in August 1963. Ahmad gained work experience as an operating engineer from George Saul Metals, a nonunion contractor.
Ahmad’s first attempt to join Local 542 was in October 1963. James Grant told him that the books were closed and no work would be available until the summer. Ahmad’s frequent subsequent contacts with the hiring hall in search of employment proved fruitless, and by spring 1964 he had accepted a job as a truck driver. Although Ahmad testified he continued to contact the union weekly throughout 1964, he received no referrals. Ahmad got his first job in July 1965 by volunteering for a three-day job and falsely representing that he had the required registrant’s book. He ultimately worked three weeks on a front-end loader and was laid off when the job ended. When James Grant, the hiring hall agent, later confronted him with his misrepresentation, Ahmad explained that it was the only way he felt he could get work. It was then that Grant accepted Ahmad’s tender of union fees and dues, issued him a B registrant’s green book, and told him he would get oiling jobs.
Between September and November 1965, Ahmad received six referrals, four of which were oiling jobs. His next referral did not come until May 1966 — a one-day job. Ahmad testified that his referrals appeared to cease at a point in November 1965, a time when William Ciavaglia replaced Grant as the hiring hall dispatcher. Ahmad filed a complaint of unfair discrimination with the Commission on Human Relations sometime during this period, but the Commission ruled against him.
Ahmad applied for the union apprenticeship program on December 27, 1965, but was told by Robert Emriek, the apprenticeship coordinator, that he was ineligible because he had previous experience on operating engineers’ equipment. Emriek then stamped a “B” on Ahmad’s registrant’s book, designating it an operator’s book. In *362spite of this Ahmad continued to pursue admission to the apprenticeship program, for upgrading his skills, and took a written test. In May 1966 he was notified of his disqualification because of low scores.
On July 3, 1966 Ahmad was again contacted by James Grant, then the C Branch agent, and told to report to the Union Paving Company. Grant informed Ahmad that his affiliation would be with the C Branch; Ahmad’s acceptance was under protest that as a B registrant he wanted work commensurate with his operating engineer branch status. For the next two years Ahmad did no operating work, but did non-construction C branch type work, earning between $2.50 and $2.78 per hour. During that time Ahmad also accumulated over 3,000 hours. Because of this, Grant verbally informed him that his branch affiliation would have to change formally to C. Again, under protest, Ahmad acquiesced. . He stated he was never informed of his eligibility to transfer from C Branch into the parent body of Local 542.
Ahmad participated in three training programs while in the C branch. Grant informed him of Operation Stepping Stone although Ahmad received no written notice. Ahmad completed the program on April 13, 1969, having received training on the back hoe. Ahmad also trained at the union training site in Delaware for a period of one month in both 1970 and 1972, learning crane operations.
After Operation Stepping Stone Ahmad received no union referrals for backhoe or crane work. Under Grant’s encouragement, he finally accepted a job as a cement mixer with Warner Company at $3.36 per hour. He later testified that while at Warner he worked in five different shops, and from his experience, only whites were crane operators or front-end loaders. Mixers, he testified, were black. As of November 1976 he was still employed at Warner in the same capacity, but earning $8.00 per hour. Ahmad indicated that since his work with the C branch was permanent, he had decided against leaving, having relinquished hope of obtaining work as an operating engineer.
(d) John Dent
John Dent received his initial training in heavy equipment operation during three years of military service ending in March 1963. In January, 1964 he enrolled in the Associated Heavy Equipment School in Florida, completing a six-week combined crane and diesel mechanic course. He then moved to Philadelphia in March, 1964, seeking work as an operating engineer, and discovered that hiring practices required that he proceed through the union.
Dent’s first contact with Local 542 was in late March or early April 1964, at which time he was informed by James Grant that no work was available and that he should return in the summer. He was given no information regarding the operational procedures of the union or its branches, nor was he given an application form, despite his stated intent to apply for membership.
Over the following two years Dent made numerous unsuccessful attempts to gain entry into Local 542. Finally, with no apparent indication of why his request for entry to the hiring hall system was accepted, Dent was permitted to become a green book registrant on April 1, 1966. He paid $30.00 in dues and completed an apprenticeship training application, leaving it with Ciavaglia and Grant. Dent, however, never participated in the apprenticeship program. Although Robert Emrick, the apprenticeship coordinator, had informed him sometime during the summer of 1966 that he was listed and would be notified, he was informed seven months later in November 1966 by Emrick’s secretary that he was not on the list. In a follow-up inquiry in early 1967, Emrick informed Dent of his ineligibility for the program based on his green book registrant status. (As indicated below a number of white B registrants were admitted to the RA program.) Dent also did not receive his first job referral until November 1966, seven months after his registration. This job was with a rental company and Dent was laid off after one week; his second referral was not until June 1967. *363All in all, during his green book registrant period, ten of Dent’s fifteen referrals were rental company jobs (short-term work) operating pippin back hoes. He received no referrals for the high paying crane work. He was never officially informed of the upgrading program Operation Stepping Stone. After becoming aware of its existence through the Labor Department and from other journeymen members of 542, Dent subsequently applied and completed the six-week program in February 1969.
In December 1969 Dent became a B branch member, having completed the requisite 2500 hours.24 Even with B branch status, about 90% of Dent’s work was limited to the pippin back hoe — there were no referrals for cranes and other heavy equipment — and these jobs were primarily rental company referrals. Dent also discovered that whites below him on the out-of-work list were being called ahead of him. He testified that he had checked the out-of-work list positions of several whites whom he observed on jobs for which he was qualified and discovered after examining the lists that they occupied a list position beneath his. He also witnessed a situation in which a black, ostensibly laid off due to the unavailability of work, was replaced by a white worker on the next day.
During Dent’s association and membership with Local 542, dating from June 1967, his jobs ranged in length from one day to several months, although short-term jobs were the norm. Dent was laid off either when there was no need for his equipment or when his jobs were finished.
The record also indicates five instances between June 1968 and July 1974 when Dent refused work referred to him by the union, and one instance where he quit a job in July 1972. Dent’s refusals to work were based alternately on his objections to the distances involved, the brevity of the work during cold weather, and his lack of familiarity with the equipment used. His quitting incident was precipitated by a disagreement over work rules with the job contractor.
On June 20, 1972, Dent was one of those attacked by a group of white operating engineers at the District I hiring hall in retaliation for the filing of the present suit. He recounts being treated poorly on his job afterwards. He testified that he was given particularly undesirable work, such as a summer job in the “hole,” 40 to 50 feet underground, breaking concrete with a hydraulic hammer. Two years later, in August 1974, Dent and his family left Philadelphia to relocate in Georgia, because of what he felt to be intolerable working conditions existing in Philadelphia.
(e) Elijah Dukes
Elijah Dukes began his operating work in the 1950s. The first union contractor who employed him was the Plumbing and Heating Service Co. At laborer’s wages, he operated a pippin for about five years. In 1958, Dukes was told by a 542 business agent — a Mr. O’Donoghue — to leave his equipment. He did not hold a union book or a registrant’s book. His immediate replacement was a white employee. The business agent told Dukes that in the next spring he would be taken into the Local. But Mr. Grant refused even to give Dukes an application. (At some time Dukes did file a card with 542 but was never contacted and did not know the significance of the card.) Once, twice, or more per week Dukes returned for a time to the 542 hiring hall in vain.
Dukes moved on to another union construction company but was still unable to enter the union. On about three more occasions he was removed from an operator’s job at the behest of a union business agent. This time such removal resulted merely in reassignment to another pippin. All in all it took Dukes eight years, until August 1, 1966, to obtain a registrant’s book.
(f) Charles Iseley
Charles Iseley, after successfully completing in 1966 a six-week privately operated training course for operating engineers, *364sought entry into the union. He was not aware of the existence of the registrant or apprenticeship programs, and despite conversations with Ciavaglia was never advised. Iseley’s attempt to enter the operating engineers profession was met with advice from Ciavaglia to obtain two references. He could not obtain such references at the hiring hall after repeated visits. Although Ciavaglia advised him to see A1 Holland, Ciavaglia did not give him Holland’s phone number. He admitted that Holland was difficult to find. On one occasion he obtained work from a union employer, but when three men asked him whether he had a book and he responded negatively, they asked him to leave the equipment. He refused, but was dismissed several days later. Thereafter he was employed in the Philadelphia prison system.
Much later, responding to a newspaper notice regarding an Outreach Program, Iseley met A1 Holland. Thereafter in 1971 Iseley applied for and took a written and field test for classification as a B registrant. Although he passed the written test he failed the field test which related to front-end loaders. Admitting his failure on this piece of equipment, Iseley observed: “I guess I may have been a little rusty after ’66 to that point.” Tr. 6387. He then became employed at Community Legal Services as an investigator and has never since worked as an operating engineer.
(g) John Dodson
John Dodson attained experience operating cranes, bulldozers, and other heavy equipment in the Philadelphia Navy Yard and several other places, and in 1954 was employed by Cheltenham Township,, a suburban community north of Philadelphia, as a dump truck driver and then as a heavy equipment operator. Because Cheltenham had a contract with 542, the union business agent, Robert Helman, objected to his working on equipment without a “book.” Dodson sought a book only through his employer, not directly from the union. Even after several years he had not received one.
Dodson later worked for Marino Massi as an equipment operator. Massi’s company was a union company, but Massi was aware that Dodson was not a member and had no book. Two business agents removed Dodson from his job a total of about seven or eight times between them. In about 1960 Dodson went to the hiring hall requesting an operators’ book and was denied one because, in the recounted words of union official James Grant, “We got too many men out of work ourselves.” Tr. 7049. On another occasion in 1967 or 1968, Dodson took to the hiring hall a note from Mr. Massi saying he owned two pieces of equipment, on the theory that it would help him obtain a book. Grant told him that if an operator were put on both pieces he would consider Dodson’s request. Dodson then related: “So, I said, what am I supposed to do? He said, well, go drive a truck.” Tr. 7051. Shortly thereafter, the dates are somewhat confused, Dodson removed to New York where he did receive a parent body book from Local 15. Subsequently he returned to Philadelphia and spoke to Grant who told him 542 needed operators. Sometime in 1968 he received a book and began receiving referrals. Although Mr. Walsh wished at first to give Dodson a B registrant book, Dodson insisted on and ultimately received a parent body 542 book on the ground that he was a parent body member of Local 15.
(h) Lloyd Hudson
Lloyd Hudson is a Vietnam veteran, honorably discharged in 1966. In order to begin a career as an operating engineer he attended a Training Services Institute program in Northeastern Pennsylvania from April 1968 to September 1968. This program has no apparent connection to Local 542 or any other defendant, and Hudson became aware of it through the Commonwealth’s Bureau of Employment Security (BES).
Having successfully completed his six month program, Hudson went to the union hiring hall and spoke to A1 Holland, showing him his training certificate. Holland said he would be in touch with Hudson “in *365the near future.” Tr. 6400. Holland took Hudson to Emrick who also said he would be in touch. With the exception that Hudson subsequently encountered Holland at the hiring hall and was told that nothing was available, no one from the union ever called or contacted him. (Hudson did make a phone call to the union and spoke to some unspecified individual who referred him to Holland. Although Hudson was told that he “first must be a member of the Local,” Tr. 6403, it is not positively clear that Hudson ever told that person that he was inquiring into anything but membership in the union.) At no time, however, was Hudson ever advised of the existence of a JATC, registrant or C branch entry.
On September 23,1968, after receiving no satisfactory communication from the union, Hudson took (at the suggestion of Mr. Sgrow from BES), a BES test, one of the two tests used to determine admittance to the JATC program at that time. Mr. Sgrow wrote to Emrick the following day stating that Hudson scored “considerably higher than the minimum scores required.” (Emphasis in original.) Hudson, never knowing of the JATC’s RA program, did not file any application for it; nor did he receive any correspondence or other communication from the JATC. A file at JATC, however, reveals Hudson was rejected by letter of May 5, 1969, for failure to take the JATC test, of which he received no notice. Until 1970 Hudson was residing at the address he originally gave the union.
Cross-examination established that Hudson had been convicted of fraud in cashing stolen United States Government Social Security checks, a crime for which he received a three year sentence.
(i) Duane B. Johnson.
Duane Johnson received his training in operating engineers’ equipment during his service in the armed forces, accumulating four years of experience by the time of his October 28, 1969 discharge. Following a two-month program in equipment repair and maintenance, Johnson served for eighteen months in the Construction Engineer Unit in Germany operating front-end loaders, back hoes, graders, and dozers. In 1967 he was sent to Vietnam as a heavy equipment operator in the 15th Combat Engineers Unit, operating the same equipment, including cranes, oftentimes in the combat zone. Returning to the United States, he was assigned to the 75th Construction Engineers at Fort Meade, Maryland, again as a heavy equipment operator handling front-end loaders, dozers, graders, backhoes, scrapers, compactors, rollers and chañes.
Johnson’s first attempt to join Local 542 was in September 1969 when he was advised by Ciavaglia to reapply after his discharge. Johnson returned within a week of that date, and was informed that his military service was not suitable for civilian purposes. Ciavaglia told him that further .training at a heavy equipment operating school would be necessary. Johnson was not told of the union’s registrant program, procedures for obtaining a green book or taking a field test, the apprenticeship program, potential C branch membership, or the necessity of paying a service fee for work referrals.
Johnson borrowed $1,000 to cover the cost of attending a two-month program at the American Training Services, Inc., in New Jersey. He trained on the identical equipment he had used in the army. Upon completion in February 1970, Johnson returned to Ciavaglia and was told that because of the large number of unemployed, it would take at least one to two years for him to obtain work, but he would be put on the out-of-work list. Johnson continued to call or visit the hall weekly, but was unable to procure any jobs from Local 542 through the end of 1970. Johnson was informed at this point of the RA program and he applied for entry thereto toward the end of 1970. He was never notified concerning that application. Between October 1969 and December 1970 Johnson worked as an operating engineer on two nonunion jobs. In February 1971 he accepted a nonunion engineering job with I.T. & T. in Greenland where he worked until the end of 1971.
*366Johnson’s third but unsuccessful attempt to join Local 542 followed his return to Philadelphia from Greenland at the end of 1971. At the beginning of 1972 he applied a second time to the Apprenticeship Committee, and at that time met Emrick. At the end of their conversation, Emrick gave Johnson his card, indicating that he should present it to Ciavaglia, the dispatcher. Ciavaglia refused to acknowledge the gesture, and Johnson was still without work.
In May 1972 Johnson was tested on the front end loader, back hoe, and bulldozer as part of the union field test. He passed; this was the same equipment he had used in the army. He later spoke with Ciavaglia who again indicated that high unemployment had virtually foreclosed present and near-future work opportunities for him. Johnson was not given a green book, and he was advised by Ciavaglia that it was unnecessary to pay dues since it would take at least two to two-and-one-half years before he would obtain work. He was told that he would be placed on the referral list. On cross-examination Johnson acknowledged he had on February 17, 1972, signed registrant papers which contained notice of referral fees, but he said that he did not pay the fees based on Ciavaglia’s statement to him. He never received any subsequent notification from the union that his registration card had been withdrawn for any reason.25 Concerning his RA application, Johnson was notified that because of changes in federal regulations he would have to reapply. He was told this after having successfully achieved a waiting list status. He had not received any response to his reapplication.
Johnson made other unsuccessful attempts to get union work during the summer of 1972 and throughout 1973. Finally, on August 30, 1973 he filed a charge of discrimination with the EEOC. One month later he received a call from Ciavaglia asking if he was available for work. Johnson testified that when he found out it was for “one or two days” he refused acceptance because he then held a steady job. As of November, 1976, Johnson has had no additional offer of jobs through Local 542, although he indicated that he still wanted to be an operating engineer and a 542 member.
(j) George Benjamin
As a young boy George Benjamin operated machinery on his father’s farm in South Carolina. By the time he moved to Philadelphia at age eighteen, he had operated tractors, dozers, cranes and front-end loaders. For several years he worked as a general laborer for wrecking companies, and in 1962 was hired by Hauser. Within a few months he was promoted to foreman, responsible for moving (not operating) heavy equipment on Hauser’s union jobs. In 1963 or 1964 he began operating heavy equipment-including crane loaders, small dozers, and pans — on a nonunion job at a Hauser dump. Benjamin worked there through 1973 when he left to start his own demolition business. Never financially successful at this venture, he went to work for Hargrove Company, doing nonunion work on crane loaders and backhoes. As of November, 1976 he had returned to Hauser, again moving heavy equipment.
Benjamin attempted to join the union in the spring of 1969 and spring of 1970. Both times he was told by Ciavaglia that the books were closed and no applications were available. On his third attempt in 1971 he was given an application and later notified of his eligibility to take the registrant test (written and field). Benjamin passed the tests on October 21, 1971, and was told he could obtain a registrant’s book at the hiring hall. On November 2, 1971 Benjamin paid $49.00 in fees and was issued a green book. He testified that Ciavaglia and Holland told him he would be notified for work, and relying on that did not actively pursue employment by regularly contacting the union. Benjamin’s first job offer came by *367letter almost two years later, arriving while he was in North Carolina visiting his mother who was ill. His wife phoned him this information and he returned to Philadelphia that night, visiting the hiring hall in the morning. Ciavaglia had already dispatched another operator saying he could not reach him.26 No other jobs were ever offered to Benjamin.
In addition, in spite of Benjamin’s timely remittance of dues throughout the period from November 1971 through November 1976 (totaling $500), his name was omitted from the Group III out-of-work list from December 1978 through June 1975. In addition, testimony also indicates that although his name was on the July 1973 list, it was crossed off and he was not advised of that fact. (There is no testimony indicating whether he was excluded from the August through November 1973 out-of-work lists.)
Benjamin’s most recent union contact was in October 1976 when he spoke to William Sautter about his situation. Sautter requested that he fill out a work sheet, and then advised him that the employment situation was bad but that he would be notified of the availability of work later.
(k) Timothy O. Roundtree
Timothy 0. Roundtree became a green book registrant of Local 542 on January 2, 1970, after two previous unsuccessful attempts to gain entry. His first attempt was in October 1968. An unidentified woman at the union office informed him that the books were closed and that she could not help him. Roundtree’s second attempt was in July, 1969, following a reprimand by Holland for handling operating engineers’ equipment for Tony DePaul without union clearance. Holland directed Roundtree to contact Ciavaglia for union registration; however, when Roundtree approached Ciavaglia with Holland’s message, Ciavaglia disclaimed any knowledge of that contact. Following a second reprimand in later December, Holland finally agreed to meet personally and register Roundtree at the Union.
Roundtree paid his $27.00 dues and his application was accepted despite Ciavaglia’s protest that his entry be delayed until spring. At that time Roundtree was not provided with either a green book or a copy of the union contract. He was also not informed of the hiring hall procedures, including the 2500 hour work requirement for union membership, although he was warned against operating any equipment without prior union approval.
At the time of his registration, Roundtree was employed by Tony DePaul for whom he had been working since 1960 as a laborer and truck driver. It was there that he gained experience handling operating engineers’ equipment. However, after Round-tree became a registrant, DePaul requested the union to clear Roundtree for work as an operating engineer. This request was refused. Roundtree thereafter continued to work for DePaul in a non-operating engineering capacity until he was laid off in December 1972.
Despite his status as a green book registrant during the period between 1970 and 1976, Roundtree never once obtained work as an operating engineer — this despite the fact that he remained constantly available, tendered timely payments of his union fees (totaling $700.00) and pursued employment by regular calls and visits to the union. On one occasion, as late as January 1976, when he complained to Ciavaglia and Walsh about his lack of work, Roundtree was told that times were bad and many persons were without work. However, several months earlier, during a June 1975 visit to the District I hiring hall, Roundtree had discovered that he was not listed on the union out-of-work list. He complained at that time to Ciavaglia’s assistant. Exhibits admitted at trial showed that his name did not *368appear on out-of-work lists from December 1973 through January 1975, but that on the Group III list for the combined months of January through June 1975, his name did appear in handwriting above the other typewritten names.
The cross examination of Roundtree focused on the issue whether Roundtree’s name was omitted from the work list due to his failure to notify the union of his change of address. Roundtree’s testimony is that he did advise the union of the change of address in October 1972 when he paid his dues by mail (that notice was sent accompanying his dues book and payment). The union’s position is that it did not receive notification until April 2,1975 when Round-tree informed Ciavaglia of the change during a visit to the hall, as documented in his work record. The union thus inferentially attributes the omission of Roundtree’s name from the list to his failure to respond to a letter of August 16, 1973. Roundtree, on the other hand, testified that he never received the August 1973 letter because it was not sent to his new address, and that the union had been notified of that change in October 1972. Roundtree supports his position by testimony that the union has continued to return his dues book to his new address with each subsequent payment he has made since October 1972.
(1) Cleveland Allen
Cleveland Allen accumulated six years experience operating heavy equipment while serving in the armed forces from April 1965 through March 1971. Following completion of a three month training program in the U.S. Corps of Engineers where he learned to operate dozers, bulldozers, rubber tire pans, trench machines, and air compressors, Allen was sent as an operating engineer to Vietnam twice (and received three commendations there). He was also sent to the Korean DMZ, and assigned to several stateside engineering units working with civilian operating engineers. Soon after his discharge on March 24, 1971, he further supplemented his training by attending the Universal Heavy Construction School.
In June 1971 Allen was referred to Local 542 by a state employment office. At the hiring hall he spoke with William Ciavaglia who provided him with an apprenticeship application. Several weeks later he was notified by letter that he was to take the union’s written and field examinations on August 28, 1971. Allen passed both tests and became a green book registrant that day. He was never accepted in the apprenticeship program. His first job referral as a grader operator was with J. D. Morrissey on August 31, 1971. He testified that although he had worked satisfactorily as a grader operator for the first three weeks, he was unexplainably replaced by a white operator and reassigned to a lower paid roller position. On November 22, 1971 when he was laid off, he noted that the operator who had replaced him on the grader was still working. Allen also testified about his referral to the Industrial Construction Company as a grader on June 8, 1972. When he reported to work he was assigned to a lower paying roller position and, when he called to complain, was then told by Ciavaglia to keep it. Allen indicated that these switches occurred frequently in his experience.
Allen’s last job working as an operating engineer was in October 1973 with Union Paving Co. Following that, he worked as a laborer for Buchtel Power Company at its Lynnewood Plant from August 27, 1974 through October 27, 1976 when he was laid off.27 He once again contacted Local 542 in an attempt to get work, but was told by William Sautter that as a Group II out-of-work registrant 700 Group I out-of-work operating engineers were ahead of him.
Allen testified that he had never refused a job from the union and had always been available for work as an operating engineer; that he had never been fired or sus*369pended; and that he had always tendered timely payment-in-full of his union fees. In 1971 Allen worked 522 hours; in 1972, 689 hours; and in 1973, 575V2 hours.
(m) Conclusion
The above individual narrations reveal in varying degrees a pattern of delay or diversion either in acting upon or in entertaining minority applications for hiring hall status as well as a pattern of neglecting to advise these minority applicants or would-be applicants of the available methods of entry. Each of the above individuals had the requisite qualifications to obtain entry through at least one entry method. The resistance to entry of these individuals was not based upon a view that they lacked such qualifications. In most of the cases, years passed before any access was obtained, and once registrant or other hiring hall status was achieved, the referrals to operating engineer jobs were few. Some of the above individuals were replaced on the job by white operators without explanation or were deleted from the out-of-work list without cause.
I find the above narrations credible in their description of resistance to equal minority opportunity through the hiring hall. The testimony of Fox, Ahmad, Dent, Dukes, Iseley, Dodson, Hudson, Johnson, Benjamin, Roundtree, and Allen persuade me that union officials acted deliberately to delay and resist applications, to fail to advise applicants of methods of entry and application procedures, and to fail in some instances to include on out-of-work lists those entitled to be listed. This conduct is consistent with the kind of deliberate resistance to minority participation shown by the deceptions contained in the Affirmative Action Program Agreement, replacing the Philadelphia Plan, and in the treatment of Benjamin Franklin trainees and other acts detailed above. This individual proof in turn supports the statistical and other proof of intentional discrimination (although such support is not here deemed essential) much as did the evidence in Teamsters that
“[njumerous qualified black and Spanishsurnamed American applicants who sought line driving jobs at the company over the years, either had their requests ignored, were given false or misleading information about requirements, opportunities, and application procedures, or were not considered and hired on the same basis that whites were considered and hired.” [Teamsters, 431 U.S. at 338, 97 S.Ct. at 1856 (quoting district court)].
In this case the individual minority applicants above were sometimes ignored in their efforts to apply and were given misleading or incomplete information about methods of entry. Although in most instances plaintiffs have not specifically shown that the resistance to entry or employment of these individuals was followed by specific instances of unresisted entry of white applicants in comparable positions, the evidence does show that there was a constant flow of entries into the union and into the registrant program during every year in question. Union or affiliated entrants viewed in general have been, disproportionately to the labor pool percentage, white entrants. In this respect the requirements of McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), have been met, although they need not have been to have rendered the individual proof relevant. Cf. Furnco Construction Corp. v. Waters, 438 U.S. 567, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978). This individual proof is not to be viewed out of the context of the other proof in this case and I repeat that its principal effect is to illustrate through individual instances how entry and participation of minority applicants (statistically and otherwise proven to have been intentionally discriminated against) was resisted by union officials.
I. Rebuttal
Plaintiff’s suit involves claims of both a pattern of intentionally discriminatory treatment, International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977), and discriminatory disparate impact, Griggs v.
*370Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971). Under Title VII, and by analogy under § 1981 in the employment discrimination context plaintiff bears the burden of proof (persuasion) on the basic elements of his claims — (1) that intentional discrimination took place or (2) that some employment practice produced a disparate racial impact.
As to the claim of intentional discrimination particularly, plaintiff may establish his prima facie case by statistics alone, see, e. g., Hazelwood School District v. United States, 433 U.S. 299, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977), or by a combination of statistical and non-statistical proof. Plaintiff bears the burden of proving a prima facie case that a discriminatory policy actionable under Title VII or § 1981 exists. Teamsters, 431 U.S. at 360, 97 S.Ct. 1843. The burden then “shifts to the employer [or union] to defeat the prima facie showing of a pattern or practice by demonstrating that the [plaintiffs’]'proof is either inaccurate or insignificant.” Id. (emphasis added). It appears clearly that this shifting burden requires a defendant to persuade the court of inaccuracy or insignificance: if defendant produces no evidence, plaintiff necessarily prevails; if defendant produces some evidence, it must be forceful enough to persuade or once again plaintiff prevails.
The burden of proof issue is slightly different where only a disparate impact claim is being considered because in such a case the defenses may be different. A defendant may seek to demonstrate inaccuracies or defects in the calculation of disparities or he may seek to demonstrate that those disparities are permissible under Title VII because of “job relatedness,” for instance. E. g., Albemarle Paper Co. v. Moody, 422 U.S. 405, 425, 95 S.Ct. 2362, 45 L.Ed.2d 280 (1975). In .the latter circumstance the defendant bears the burden of persuading the court that there is “a manifest relationship to the employment in question.” Id.
This somewhat complex structure is made much simpler in the circumstances of this case because the plaintiffs’ case on a preponderance of the evidence, statistical and otherwise, is persuasive both that the union practiced a pattern of intentional discrimination and that union practices in the overall operation of a hiring hall for operating engineers created substantial racial disparities. Defendants’ rebuttal, hence, has on the whole simply failed to render plaintiffs’ proof unpersuasive and a fortiori has fallen short of disproving plaintiffs’ case on a preponderance of the evidence standard. Some of the major features of defendants’ rebuttal evidence will be summarized below.
1. Experts Statistical Analysis
a. Labor Pool
A basic aspect of the plaintiffs’ statistical proof was the gross statistical disparity calculated by Dr. Siskin between the minority composition of the operating branches of 542 and the minority composition of the labor pool consisting of males between the ages of 18 and 65 within the jurisdictional reach of 542. Siskin found the labor pool to be 11.5% minority, whereas union minority composition, as of 1971, was 4.4%. Siskin found the disparity to be significant at below the 10-23 level.
Dr. Michael L. Wachter, one of the union’s expert witnesses, is a Professor in the Department of Economics at the University of Pennsylvania. Dr. Wachter is described in Local 542’s Brief as a labor economist. It was his opinion that the labor pool of minority individuals should be defined much more restrictively than Siskin defined it.
Wachter in essence believed that there were numerous factors (cultural, socioeconomic, religious, vocational, educational, etc.) which would render inaccurate a calculation of labor pool based simply on the number of male minority individuals aged 18-65. Wachter theorized that these factors may tend to make minority individuals, for whatever reasons, less likely to possess skills needed to enter the union, to succeed in the union, or to be oriented in the first place toward operating engineer work, etc. Wachter, however, provided no analysis *371which sought to quantify the effects of these factors. He plainly did not offer any evidence that minority individuals disproportionately lacked the skills or availability to do operating engineer work.
Apart from Wachter’s failure to give the court an indication of the strength of the effects of the factors he referred to, there is other evidence in this ease indicating that his hypothesis is in error. Wachter’s argument, synthesized, is that assuming arguendo that the union practiced no discrimination against minorities, the minority membership in the union may still be much less than 11.5%.28 This, his theory goes, might be due to pre-labor market factors such as an inability of minority applicants to gain admittance for lack of qualifications or a lack of motivation to seek operating engineer work.
With respect to his hypothesized lack of skills among minorities, it must be pointed out that there have been and to some extent still are entry methods requiring no prior skill or level of educational attainment. Indeed, only the RA program, recently, has required a high school diploma. The A registrant program required no skill or experience. An A registrant could progress to union status by accumulating hours in essentially unskilled areas. The evidence also indicates that the B registrant program which required operating engineer experience and, after 1970, required testing has held a greater percentage of minority individuals than did the A registrant program. And it is to the B registrant program which elsewhere in his report Dr. Wachter himself points for the conclusion that minority individuals are represented in relatively high percentages and tend to be more successful in reaching membership status. The C and D branch organization and C miscellaneous methods themselves require no particular skill or experience to do operating engineer work, and transfer from these branches to the operating branches can be effected. Certainly unauthorized methods of direct entry require no level of skill or experience; and the RA program, once it was extended to those who were not union members, was particularly directed at unskilled A registrants. Given the existence of these entry avenues, one might reasonably expect that societal deprivations bearing upon skills, if evidence showed their existence, would over time have no effect or little effect on composition.
To the extent that Dr. Wachter theorized that the actual number of available minority applicants is less than Siskin’s estimate because of a lack of minority motivation, disposition, or orientation toward operating engineer work, he is contradicted by the numbers of minority applicants to the RA program, particularly in 1969 and after 1972 when federal regulations required examination of minority underutilization. The percentage of minority applicants typically exceeded the 11.9% minority goal which the JATC itself established in fulfillment of its obligation to calculate the labor pool.29
Siskin’s 11.5% minority labor pool definition for the above reasons withstands Dr. Wachter’s criticisms. More importantly, in the absence of any proof that in fact there is a relative lack of skill among minorities to become operating engineers a reduction of the labor pool size is wholly inappropriate.30 Further, given strong indications *372that particular levels of formal education are not necessary for effective performance of operating engineer tasks, to impose an educational qualification on the labor pool definition would not only contradict the facts, but would indirectly insert into plaintiffs’ analysis a non-validated qualificational standard which, by Wachter’s theory, could have the effect of disproportionately excluding minority individuals from the composition of that labor pool. This in itself seems impermissible absent any concrete evidence.31 It is intriguing to note that Dr. Wachter suggests that the black labor pool should be smaller for the purposes of this suit because percentagewise fewer blacks have high school degrees, yet he seems oblivious to the fact that all of the members of the union’s Executive Committee had failed to complete high school. Bluntly, Local 542 is not seeking to train applicants for the faculty in the Department of Economics at the University of Pennsylvania, and Dr. Wachter’s restriction of the black labor pool fails to pass even a common sense standard or one grounded on a rational statistical method. Wachter’s assertion that the lack of educational qualifications is a signal of a lower level of operating engineer related skills is unsupported. Upon any comparison, the testimony of Dr. Siskin was far more credible and persuasive than Dr. Wachter’s on this issue.
There are a few other theoretical factors mentioned by Dr. Wachter as bearing upon labor pool definition. Defendant Local 542 has chosen in its brief, however, to discuss only those addressed above. These additional factors will be noted only briefly. Wachter asserts that the geographical consideration of the labor pool cannot be limited to 542’s jurisdiction, that the intermittent nature of the work limits the numbers of those willing to be operating engineers, and that becoming an operating engineer is a matter of self-selection and individual choice. These theoretical assertions do not alter my view of Siskin’s labor pool analysis. Suffice to say that Wachter does not cite to any evidentiary material at all to suggest that these are in any way significant factors which would disproportionately affect the minority labor pool.
Dr. Wachter, in addition to speculating about the above factors (including education and prior training or skill), has referred to several other considerations whose effects he sought to calculate. Specifically, he sought to quantify the effects of considering “labor force” (rather than simply the population of minority males between the ages of 18 and 65), possession of an automobile and telephone, and the residential pattern of current members.
Wachter concludes that it makes no sense to consider persons in the minority labor pool as defined by Siskin who are not available for work. Thus he limits the labor pool to those in the “labor force,” excluding the persons identified in the Census Bureau Statistics as not actively seeking employment and thereby reducing the minority composition of the labor pool to 10.1%. There is much superficial logic in this reasoning; however, Dr. Wachter is not precise *373as to the reasons why someone may not be in the labor force. It is possible that some of those not actively seeking employment are discouraged by the unavailability of work or by what they might perceive to be discrimination in some areas of employment. (Dr. Wachter himself referred to high unemployment in explaining the high rate of minority applicants to the registrant program in 1970-1973.) A greater percentage of this group may be minority individuals. If jobs were to become available, this segment might well become a work force segment even in Wachter’s terms. The crucial point is that it simply cannot be said from Wachter’s analysis, insofar as it was developed, that there would result a disproportions te diminishing effect on the minority labor pool, considering both the labor force figures and that segment which would join the labor force if jobs were available. Given the limited extent of the proof and its arguable nature, I reject the labor force reduction which Wachter suggests. Wachter’s argument does not provide a tangible basis for seriously questioning the appropriateness of Siskin’s labor pool definition.
W achter further proposed that the minority labor pool be reduced by eliminating those who do not possess an automobile or a telephone. An automobile, Wachter maintains, is a practical necessity for an operating engineer. Likewise, he believes a telephone is essential for participation in the union’s referral system. Wachter is correct to the extent that a phone and an automobile could be and probably would be useful to an operating engineer. But in practical terms his analysis obscures an important reality. Common sense dictates that there are many alternatives to possession of an automobile or telephone. There are other methods of communication of referrals from the hiring hall, such as, communication through a friend or visiting the hiring hall once a job is completed or, perhaps simplest of all, obtaining a telephone. Likewise, an automobile can be borrowed, friends or relatives may provide transportation, possibly other forms of transportation may be used, or an automobile may even be purchased after entry to the hiring hall is achieved. In short, Wachter’s analysis on these points seem simplistic, absolutizing characteristics which can be altered or adapted with relative ease. It seems patently improper therefore to use such characteristics to rule out employment through the union or define a labor pool on that theory. Indeed, the union itself does not require possession of an automobile or phone.
Wachter’s final empirical adjustment of the minority labor pool is based upon what he perceives as an appropriate alternative method of calculating that pool. Wachter suggests the minority labor pool can be defined by determining the residential patterns of current members “to define the expected number of minorities to come from each area.” Wachter’s use of this approach creates grave doubt whether he understands the purpose of statistical evidence in an employment discrimination case and causes me to question further the reliability of his conclusions. It is hard to imagine that an expert witness seeking to determine the percentage of the labor pool composed of minority individuals in the context of a suit alleging purposeful discrimination (particularly in membership) would ever seek to define that labor pool by the characteristics of current members. Such an analysis obviously tends to beg the question raised by this suit and must therefore be rejected.32
Two other experts made comments critical of Siskin’s labor pool calculations. They are Dr. Lewis J. Perl, expert for defendant Glasgow, Associations and contractors, and Dr. Arthur P. Dempster, another expert witness for Local 542. Dr. Perl is Vice President of National Economic Research Associates, Inc., and Dr. Dempster is *374a professor in the Department of Statistics at Harvard University who has been Chairman of that Department, based on a rotating appointment system. Dr. Perl taught economics at the New York School of Industrial and Labor Relations at Cornell University and has had experience in various areas of economic labor analysis. Dr. Dempster’s specific area of concentration within the field of statistics is mathematical statistics.
Dr. Perl’s basic criticisms repeat some of the criticisms of Dr. Wachter. They are therefore subject to the preceding analysis. I will, however, catalog Dr. Perl’s criticisms separately.
While Dr. Perl agrees that in general, despite some minor discrepancies, Dr. Sis-kin’s calculation of the percentage of minority membership in the union is accurate, he disputes the minority labor pool percentage of 11.5% by proposing reductions through adjustments according to “labor force,” automobile and telephone ownership, educational achievement, and acquisition of skills. Perl also criticizes the temporal limitations of Siskin’s labor pool estimate. For the reasons earlier expressed, the labor force reduction must be rejected.
Unlike Wachter, Perl states explicitly that some number of those who are not in the labor force may consist of persons who are simply not interested in seeking work. In the absence of further evidence, as was said before, it is possible that this group is disproportionately minority and that its interest in seeking employment may be renewed if employment became available. Failure to try to sort out the group of uninterested workers makes Perl’s reduction speculative. Little time need be spent on the automobile and telephone ownership factors. Perl does not state that they are absolute necessities but only that participation in the hiring hall “probably requires” or “may require” these instrumentalities. For the reasons expressed above these factors are likewise rejected.
Concerning his conclusion that educational qualifications should be used to define the minority labor pool, Perl cites to the 22.3% black failure rate on the Bureau of Employment Security aptitude test administered by the JATC. This, Perl states, is “nearly twice as high as the rate for whites.” Perl then states “if we were to assume similar disparities in educational qualifications between blacks and whites .” more generally, the minority labor pool would be considerably reduced.33 Significantly, Perl did not say that such an assumption was in fact warranted. Moreover, the BES test, which he stated, matter of factly, as having been “validated,” is found in this decision not to have been validated.
Perl’s suggestion that acquisition of skills outside the union is a factor to be considered in calculating the racial composition of the labor pool fails to consider the existence of unskilled entry routes throughout the relevant time period. Further, Perl, like Wachter, points to no evidence that minorities are disproportionately unskilled for operating engineer work.
As a separate point, Dr. Perl raises the question of the propriety of looking to any static figure representing the minority labor pool at any given time because the composition of the union has developed over a period of years. Consistent with the usual approach of Dr. Wachter, Dr. Perl declines to develop any substitute statistical measure or to indicate whether such a measure would disproportionately affect minority labor pool composition. I believe the measure chosen from the most recent census data is an effective one for purposes of generating a (rebuttable) statistical conclusion. Speculation is not effective rebuttal. And although Dr. Perl observed that the entry rate of minorities in the period 1966-1971 was a better focus for comparison with a labor pool figure, he made no effort to calculate the results of that comparison.34 In fact Dr. Siskin’s evaluation of *375this entry data, discussed supra, indicates on the average a 6.4% minority entry rate during these years. This is statistically significant at below 1 in one million and in context supports the existence of discrimination, not its absence.
Finally, Dr. Dempster, testifying on behalf of Local 542, has chosen expressly to leave for the other experts the task of analyzing Siskin’s labor pool calculation in detail. However, he has voiced a general view that the general community was not a valid source for comparison, and points out that defining the labor pool may be impossible. I disagree with Dr. Dempster’s general conclusion in the circumstances of this case.
It is clear that the defendants believed it sufficient principally to raise theoretical questions about the value of Dr. Siskin’s analysis and in most instances chose to do no more. With the burden on defendants to “defeat” plaintiffs’ prima facie case by demonstrating inaccuracies or insignificance, this is patently insufficient. But apart from the burden of proof question (for this is not a case in which the evidence is at an equilibrium) defendants have simply not raised any tangible or credible reason why Siskin’s labor pool estimate should be viewed as anything less than persuasive. In light of the preponderance of the evidence standard in this case, I find in favor of the conclusions of Dr. Siskin, whose credibility has not been undermined.
b. Features of Entry
Dr. Wachter, beyond his conclusions as to the size of the minority labor pool, was also of the opinion that the statistical evidence relating to entry and entry methods failed to reveal discrimination. I will summarize Wachter’s major points on this subject.
First, Wachter takes issue with a projection by Siskin that by year 2001 minority membership would be only 7.2%. This projection was based on the trend, as Siskin calculated it, from data for the years 1969-1971. Siskin also determined, based on his calculation of the trend from data for the period 1966-71, that by the year 2001 minority composition would be 6.7%. Wachter emphasizes the unreliability of such projections as actual indicia of future composition. Wachter is, in my view, entirely correct, but he has missed the point. Siskin was not offering the projections to predict the future or to suggest that the real legal goal was to attain an exact parity in minority and white composition. Siskin qualified his projections by making them contingent upon the continuation of the same average minority entry rate and union growth rate as those for the periods 1966-71 and 1969-71. The real point is that Siskin was not saying what the future will be like, but was offering a way of looking at the evidence for the 1966 — 71 and 1969-71 periods. The “projections” thus serve only to describe the effect of entry on composition within the years 1966-1971 and certainly have no actual predictive value.
Wachter also criticizes as statistically insignificant Siskin’s determinations that the hours and wages of minority RAs were less than those for white RAs. Siskin’s analysis itself did not establish significance statistically on the relative hours and wages of each class. The only statistical significance was in the pattern of diminished hours and wages for minority RAs. This significance was at below the .05 level, the basic level at which significance begins. The statistical analysis is, however, the least important aspect of this evidence on diminished hours and wages among RAs. Although such a statistical conclusion would not in itself permit an inference of any discriminatory cause, given the other proof in this case the existence of those average disparities is meaningfully consistent with the claim of discrimination.
Turning to Siskin’s analysis of the registrant program, Wachter asserts that BF graduates must not be excluded from consideration of total registrants. Although the union bears no responsibility for the development of the BF programs (and despite the unfulfilled assurances by union official A1 Holland in the recruitment process), BF graduates did become B registrants and therefore must certainly be counted as such. (The occurrence of BF programs *376must, however, be viewed as a unique experience.) In part due to this inclusion, the post-1969 data on entries to the registrant program does not indicate discrimination when viewed alone. The 1966 — 1969 figures, which according to Siskin show a much lower rate of entry by minorities (significant at below the .01 level), are, Wachter complains, based on estimates of minority membership. Wachter does not specifically criticize their accuracy as estimates. Dr. Siskin’s Report 2, p. 17-18 (Plaintiffs’ Exhibit 168), fails to mention whether these calculations are estimates or, if they are, on what they are based.
With regard to Wachter’s criticism of Sis-kin’s calculation that the hours and wages of minority BF graduates were lower on average than those of white graduates, it must be pointed out that Siskin himself drew no statistically significant conclusions. The disparity, however, exists in real terms and hence corroborates the fabric of other proof in this case. Despite Wachter’s theory that a difference of skills between BF white graduates and BF minority graduates may explain the disparities, there is no evidence to suggest that the skills among this group of co-trainees are unequal. The suggestion simply is unsubstantiated by evidence. Inasmuch as Wachter himself points out that the rate of attaining union membership from the registrant program is slightly higher for minority operating engineers, the bald assertion that minority graduates of the BF programs might be less skilled is rendered even more suspect.35
Criticisms by Wachter that Siskin has failed, in addition to his other argued failures, to show the mechanism by which entry discrimination was accomplished are entirely misplaced. Siskin testified that he was not capable of drawing a conclusion from the statistics that legal discrimination occurred. That conclusion is properly left for the court. In providing statistical analysis, Siskin established an evidentiary base supporting the inference of discrimination and playing an important role in the achievement of plaintiffs’ prima facie case. Siskin’s expert testimony must not, however, be viewed alone. Nor is it legally necessary in establishing a prima facie statistical case that plaintiffs also explain some mechanism by which the inferred purposeful discrimination took place. Discrimination is no less illegal because it may be covert. In this case there is evidence to indicate that would-be minority operating engineers have been discouraged or prevented from filing applications and that they have not fully been advised of the variety of entry mechanisms upon inquiry. In total, plaintiffs’ proof more than adequately establishes those elements necessary to support the inference of intentional discrimination.
Dr. Wachter gives an interesting response to Siskin’s observation that the RA entrants represent a relatively small part of union entrants. Wachter’s response (which does not really direct itself to the issue of discrimination in membership, referral, and entry) is posed rhetorically:
The question is what should the union do differently? First, if the Union stopped its efforts at unionizing new workers, registered apprentices as a percentage of total new entrants would rise dramatically. No new jobs would be created, no additional minority workers would be hired, but the Union’s statistics would look better to Dr. Siskin. Alternatively or in addition, the Union could scrap the registrant system. This would also improve the relative importance of the registered apprenticeship program. [Union’s Exhibit 428 at 66-67]
There is certainly no obligation on the union’s part to forego any entry methods which do not in themselves discriminate or participate in overall discrimination. Doing so would not change the issues of liability. *377There is an inflexible duty not to discriminate. And in that respect the recent minority entry figures for the RA program are properly viewed in contrast not only to an array of other entry methods but also to the earlier RA experience. If it was Dr. Wachter’s intent to recommend a serious response to the evidence of discrimination in this case he has fallen short of his purpose.
c. Hours and Wages
In order to present the defendants’ criticism of Dr. Siskin’s determination of significant disparities in the hours worked and wages obtained between white and minority union member operating engineers, I will consider what I view to be the major observations of each of the defendants’ three expert witnesses. It must be remembered, preliminarily, that Siskin’s conclusions stem from two separate studies, the first covering the years 1969-1971, the second covering 1972. Both studies included regression analyses seeking to equalize the factors of age, group list, seniority, branch and district.36
i. Dr. Wachter
Dr. Wachter disputes Siskin’s analysis of hours and wages, characterizing it as misleading. He did not, however, perform any duplicative study which might suggest a different result. He does not dispute the appropriateness of regression analysis in general as a way of isolating a disparate racial effect in hours and wages.
Among his foremost objections is the claimed impossibility of testing the accuracy of calculations because the individual effects of the specific variables Siskin considered are not fully presented in his report. He also complains of the difficulty created by Siskin’s failure to indicate standard errors. (The standard errors would be used to determine confidence intervals indicating the strength of the conclusions as to race effects.) In a similar vein Wachter levels a general criticism that he did not have available to him the regression analyses which he assumes were used to develop the regression which Siskin ultimately reported.
Wachter also disputes Siskin’s conclusion that, in part because of the high significance level of the regression, the appropriateness of a “linear” model is confirmed. Siskin used such a model as a method of equalizing the variables of age, branch, district, out-of-work list, and seniority. And Wachter questions the failure to add other variables, particularly the variables of skill (ability on machines), pre-labor market lack of endowment, and the self-selectional aspect of acceptance or refusal of referral.
Wachter’s criticism relating to the failure to provide underlying data is not a criticism of Siskin’s conclusions or of his analysis, but of the difficulty with which Wachter was presented in attempting to parse it. Because Dr. Wachter had considerable information on Siskin’s mode of analysis he was certainly not left defenseless to seek to attack it. Indeed, he had full opportunity to attempt a reproduction of Siskin’s work if he thought there was an underlying defect. This he chose not to do.
With respect to the selection of a linear model, Siskin in his Report No. 7, p. 20-21 (Plaintiffs’ Exhibit 168), observed:
It is my belief based on experience and analysis of regression techniques, that a linear model is adequate to calculate a race effect, and that adding non-linear terms and interactions would rarely produce a downward change in the race effect.
In addition Siskin concluded that a non-linear regression would be inappropriate. He did, however, run a non-linear model and achieve results comparable to his previous linear regressions. Siskin's statement that the significance of the linear regression could confirm the appropriateness of a linear model was not an unqualified one. He observed:
[The] assumption [of linearity] is . tested by the overall statistical signifi*378canee of the regression. Using the standard computer program, I found that the statistical significance of the overall regression is at well below the .0001 level. This tends to confirm the assumption of linearity, i. e., that there is a linear aspect to the relationship between the hours worked and the variables in question. [Emphasis supplied.]
Overall significance, as I understand, does not itself prove the linearity assumption.
Wachter’s remaining contention seems more serious. Siskin did decline to include the factor of skill in his 1969 — 71 regression. This could have an effect on the ultimate result of the regression if skill was indeed a significant factor affecting hours and wages and if it varied to some extent according to race. I have rejected the use of the skills factor as a theoretical basis for calculating labor pool; but it may be a slightly different question whether skills are appropriate for consideration in the regression studies. Nevertheless, there is evidence to indicate that referrals based on skill were not frequent. There is no evidence at all to indicate that skills vary by race, or that they are not fully accounted for by the age, seniority, work list, branch and district factors. Siskin based his exclusion of skills additionally upon his view that the union data on skills were “totally unreliable.” An analysis of the union records demonstrates that even the union hiring hall agents did not consider skills listed on the work records (the first page) as accurate. Time and time again there was testimony that hundreds of persons were assigned to jobs that required skills which were not noted on their records. As to the skills of the members, the testimony demonstrates the pervasive inaccuracy of the union records. Yet Dr. Wachter complains about the failure of Dr. Siskin to consider skill data which was never available in any accurate form and which his client made no attempt to correct so that Dr. Wachter could by empirical study substantiate his thesis. Further, to the extent skills may be expected to develop from experience on the job, discrimination revealed by hours disparities may have a direct effect on level of skill thus possibly rendering such a consideration inappropriate as a factor by which the racial effect might be reduced.
Given the absence of evidence as to whether skill is variable according to race, the consideration by Siskin of factors which on their face should account for skill, and the serious doubts created about the importance of skilled referrals, it is my belief that Siskin’s exclusion of that factor does not undermine the value of his determination of a significant hours and wages disparity.
ii. Dr. Perl
Dr. Perl is much more cautious and analytical about his conclusions on the hours and wages disparity than was Wachter. He repeats a number of Wachter’s criticisms, and concludes with an overall, but qualified, negative view of Siskin’s appraisal.
Perl, too, focuses on potential variables which Siskin did not include when he sought to equalize non-racial factors affecting hours and wages. In particular Perl would have added skill to the equation. He observed: “if skill is left out of the Siskin regression, if blacks are less skilled than whites and if skill positively relates to hours of employment, the apparent race effect [arrived at by Siskin] will measure differences in hours that are, in fact, attributable to skill.” Contractors’ Exhibit 6 at 10. Perl would also add to the regressions the factors of willingness to travel, the extent to which jobs are refused, and the initial method of entry. These factors, he believes, while not the only appropriate ones, must be considered before any conclusion about race effect can be drawn. I find that the failure of Siskin to use these factors in the regression analysis was not inappropriate.
With respect to the skill factor, Perl himself has great difficulty arriving at a basis for measurement. Ultimately he relies on a sample skill survey conducted by the union of members to determine their number of skills on pieces of equipment. The information he uses is based on self-assessment of *379skills. Apart from the arguability of the results of any such self-assessment, Perl like Wachter fails to consider a basic factor which tends to make skill, as he defined it, an inappropriate consideration having no permissible potential for reducing the race effect; if indeed discrimination in hours and wages exists, those persons discriminated against might tend to have developed experience on a smaller variety of equipment. Hence, the skills factor used by Perl is not a factor which necessarily precedes union discrimination and is not necessarily associated to race in pre-entry manner. Furthermore, Perl himself states that “the skill variables which are used in this analysis are extremely simplistic measures and do not provide an objective appraisal of the level of employees skill.” Contractors’ Exhibit 6 at 15. In replicating Dr. Siskin’s linear regression, Perl determined that the skills factor (apparently assuming that this factor did not itself stem from discrimination) worked a reduction of the hours disparity by eight hours.
The willingness to travel factor which Perl thought appropriate for inclusion in the regression analysis was also added to Perl’s regression. Given the sequence in which Perl considered it, this factor resulted in a six hour reduction in the hours disparity. However, Perl’s definition of willingness to travel seems artificial. He stated that “36 percent of blacks said they would take employment outside their own district as compared with 60 percent of whites.” Contractors’ Exhibit 6 at 14. This wholly ignores the fact that a large percentage of all minority members of Local 542 are heavily concentrated in the most populous and potentially the busiest District I. The white population of Local 542 is more evenly distributed. Perl hence apparently failed to differentiate the district by district effect on the willingness to travel. There is simply no reasonable support for the inclusion of this factor.
Perl also tried to account for method of entry in the regression, and hence reduced the race effect by three hours. Perl does not, however, discuss why method of entry is particularly different from branch location, a factor which Siskin considered. This could be an important consideration, unless Perl is really trying to ascertain pre-labor market “underendowment of skills,” a factor previously rejected.
In his regression Perl sought also to incorporate consideration of refusals of referrals. Use of this factor risks a mismeasurement because of possible differences in quality of jobs offered in referrals which could be related to discrimination. Perl does not state specifically the effect of this factor as he used it, but combines it with “interaction terms” to reduce the race effect by eleven additional hours to a 79 hour figure. Perl then suggests that there may be still other variables. His ultimate conclusion is that
“[a]ll we can reasonably conclude from these regression results is that blacks work fewer hours then whites. Whether this reflects differences in skill, availability or, as Siskin asserts, employee discrimination is impossible to determine from the results above.” [Contractors’ Exhibit 6 at 16],
The factors which Perl sought to add to the Siskin regression analysis were not shown by any evidence to be factors which are definite, significant or, even if definite and significant, separable from the discrimination alleged in this case. In any event, the reductions for skill, refusals, willingness to travel, entry method and “interaction terms” reduced the disparity from 109 hours to 79 hours. The disparity was not eliminated, and Perl acknowledged its existence based on the regression. Perl’s testimony and analysis do not provide, in my judgment, any firm basis for rejecting Sis-kin’s conclusions. I therefore maintain my view, on a preponderance of the evidence, of the validity of Siskin’s hours and wages conclusions.
iii. Dr. Dempster
Dr. Dempster’s criticisms of Siskin’s hours and wages analysis are much like those of Perl and Wachter. They are stated in general terms and rely principally on *380speculations about possible effects of factors which Siskin did not directly include in his regression. Significantly, Dempster provided no replication of Siskin’s analysis although he said that other analyses could be performed “cheaply” and “easily” and although he said that he might perform them.
Among Dempster’s principal criticisms are several which have been considered above. He stated that Siskin’s conclusion of a racially based disparity may have overlooked other considerations which, if included in the regression, might have reduced the race effect. He stated that the assumption of linearity was mistaken. I will not address these issues any further. Somewhat different in kind from the above assertions, however, is Dempster’s opinion that racial effect could not be attributed causally to discrimination. He does, however, acknowledge the existence of a racial disparity. He also adds that a prospective experimental study, by analogy to statistical evaluation of medical treatments, is the only kind of study which can support with any assurance a causal inference.
The suggestion of a controlled experimental test for discrimination reflects Dempster’s experience with the use of statistics in medical research. Dr. Dempster hence disputes the appropriateness of retrospective statistical studies in general. While this may be a meaningful criticism in some areas of medical research, it is not meaningful or realistic in the setting of an employment discrimination case. (Nor is it consistent with the use of retrospective statistical analyses in Teamsters, supra, and Hazelwood, supra.) It is simply not feasible in a case such as this to construct a controlled experiment designed to test a discrimination hypothesis during some future period. I am also persuaded that it is not necessary.
As a whole, Dr. Dempster’s criticisms do not undermine, in my view, the essential persuasiveness and significance of Dr. Sis-kin’s hours and wages analysis.
d. Conclusion
Based on the foregoing illustrative discussion of the opinions of Drs. Wachter, Perl and Dempster, and based upon their reports and testimony as a whole, I remain persuaded by Dr. Siskin’s analysis and credit his conclusions and credibility.
2. Other Rebuttal of Defendants
In its post-trial brief defendant 542 has enumerated a great many arguments and opinions of Homer Dawson, Local Union President, seeking essentially to undercut the statistical analysis of Dr. Siskin. On the whole I do not credit Dawson’s testimony. His layman’s approach to statistics suffers seriously from a lack of understanding of even the basic statistical methods used by Dr. Siskin. Suffice it to say that Dawson has not succeeded in reducing the probative value of Siskin’s conclusions or in obscuring the issues. The assertion that based on Dawson’s testimony errors of fact and mistaken assumptions rendered Siskin’s analysis unreliable is rejected.
Local 542 has also listed in its brief a number of circumstances which it asserts rebuts any inference of discrimination. Among these assertions are the following: that in referrals whites bypassed other whites as well as blacks; that blacks on some occasions bypassed whites; that minority individuals used Miscellaneous entry and Organization entry and in a few instances unauthorized entry; that many minorities in the union are related; that minorities including the named plaintiffs and plaintiffs’ witnesses received preferences in referrals; that whites also received short term job referrals; and that whites and blacks entered the union in the same ways.
These assertions suffer from a very significant defect. When considered closely, they are not basically inconsistent with plaintiffs’ allegations of discrimination. The fact that referrals sometimes involved whites bypassing whites as well as blacks or some blacks bypassing some whites does nothing to contradict the overall disparities which plaintiffs’ proof addressed and which have not been explained. Indeed, this prac*381tice only confirms that the referral system was not bona fide in structure. The fact that minority union members may have received a higher percentage of referrals in some years than their percentage in the union only makes more conspicuous the disparities in hours and wages. Calculating referrals as such does nothing, after all, to distinguish between short term, undesirable jobs and desirable ones. Nor does the fact that some whites received short term jobs provide any meaningful suggestion that the statistical disparities and other evidence of discrimination are incorrect. As to entry by minority members, plaintiffs never alleged the kinds of exclusively segregative practices which would be immediately identifiable, but instead plaintiffs alleged a slightly more subtle channeling of most of those minority individuals who did enter the union.
Similarly the existence of some cases of unauthorized minority entry or the suggestions of minority nepotism are insignificant given the total evidentiary picture including the overall gross statistical racial disparities. The decline of overall hours worked by operating engineers, furthermore, does nothing to explain these substantial and overall continuing comparative disparities.
The union procedures in general constituted a motley fabric of arbitrary departures from the rules. I do not find that union’s intentional discrimination was practiced in such a way as to exclude absolutely all minorities from entry or to deny absolutely all minorities from the benefits of employment. The gross disparities and other evidence are nevertheless persuasive that the union practiced intentional discrimination.
J. The Case Against JATC
Plaintiffs have also complained of discrimination against minority individuals in the selection of persons for the Registered Apprenticeship program. This program is run jointly by the contractors and the union through a Joint Apprenticeship Training Committee headed by Robert Emriek, a former member of Local 542’s Executive Board and once an operating engineer. The RA program involves four years of both classes and on-the-job training. Apprentices begin as oilers in a Group I out-of-work list. This priority listing itself makes RA status desirable.
As determined by Dr. Siskin, between 1966 and 1971 there were an average of 4.39% minority entrants into the RA program. These entrants were selected from 2771 applicants, 321 (11.6%) of whom were minority members. The statistical significance of the disparity based on the average over the six years was determined by Siskin to be at below the one in one million level in comparison with the minority labor pool. The average percentage of minority graduates for this period, as indicated earlier in the discussion of entry discrimination was 1.8%. This was determined to be significant at below the one in one billion level when compared with the minority labor pool, and is far less than minority union representation. From 1972, on the minority entry data shows significant increases.
The standards for admission to the RA program were essentially modeled on the national standards developed by the National Joint Apprenticeship Training Committee. These national standards, as relevant here, imposed age restrictions (between 18-25 years); an ability requirement; and a twelfth grade educational qualification. The defendant JATC adopted the age and education requirement (either giving preference to or requiring a high school diploma of RA applicants) and instituted a written testing procedure as set out below. Although Homer Dawson, the local union’s president, testified that the Department of Labor approved the local standards, this testimony was permitted pending documentation; that documentation was never presented.
At first the union expected to select current members of A branch (oilers) based on seniority and without discrimination. The seniority standard was approved by the Department of Labor. Emriek wrote to the Department of Labor stating that if applicants were to be selected beyond the pool of *382650 oilers, then procedures in accord with state regulations on discrimination in such programs would be developed. The approval of the Department of Labor was thus circumscribed. The standard ultimately adopted by the union, as revealed in the minutes of a JATC meeting, was a blend of “previous experience and qualifications.”
In the first year, 1966, four classes of RAs were chosen. The third and fourth year classes were chosen from A branch, and the first and second from A-registrants (non-members). Mr. Dawson testified that A-branch entrants were placed in the RA program, but were not RAs. The union, in evaluating entrants who were A-registrants, assigned point scores in five areas: educational background, age, test scores, on-the-job experience and interview results. Of particular note, one point (of a possible 20) was to be deducted for each year short of 12 high school years and for each year in excess of age 25. A' minimum score of 70 was established. The written tests employed were the Department of Navy Equipment Operators’ Exam, and a mathematical and verbal test. A-branch members were exempt from all the above specific requirements, except possibly the Navy test.
In 1967, the Navy test was replaced with tests from the Psychological Corporation. In 1968 the Bureau of Employment Security test (“BES”) was used and the high school education requirement and the age requirements were made inflexible minimum qualifications. Another test, a JATC test, was administered to those meeting all other standards. (Notably the national standards did not require any particular test, nor did they require an interview.)
These were the procedures which were supposed to be followed until changes in 1972 pursuant to federal regulations requiring RA programs to measure minority participation. 29 C.F.R. § 30.4(d)(1) (1972). Although the local JATC, along with the national JATC, objected to such regulations, they proceeded to analyze their minority situation. The JATC determined, doubtless from the standards set forth in subsection “e” which included labor market representation, that the percentage of minority entrants they were obliged to achieve was 11.9% of total entrants (not far from Siskin’s determination of 11.5%) and the committee set its selection goals accordingly. In 1972, 1973, 1974 and 1976 (no class entered in 1975) minorities entered the RA program at a rate commensurate with or in excess of their representation in the relevant demographic labor pool.
The data on minority entries into the RA program has been set out above in the discussion of statistical evidence of entry discrimination. It will not be recounted here. With respect to JATC selection between 1966 and 1968, as detailed previously there were specific instances of disparate treatment of minorities. John Dent and Willie Rush both were rejected allegedly because they were B-registrants although some white B-registrants were accepted; Robert Ahmad was rejected for failing to appear for an interview although he received no notice of the interview; and Eugene Cobbs was rejected ostensibly because he was overage at 45, although in at least one instance an even older white applicant, Clifton Killinger, was admitted and although as a union member Cobbs was not supposed to be subject to the age limitation.
Between 1966 and 1968 the principal reason for rejection of minority applicants to the RA program was that they “applied in error.” Forty-one percent of the minority applicants were rejected on this ground as opposed to 13.8% among white applicants. Dr. Siskin found this difference to be statistically significant at below the .01 level. With the advent of the BES test in 1968 and for the period 1968-70, it is more than coincidental that minorities were rejected in disproportionate numbers on grounds of failure to pass the test but very few applicants were rejected for having applied in error. The disproportionate impact of the BES test was acknowledged by Mr. Dawson. From 1968-1970, minority individuals failed the BES test at a 23.24% rate, whereas whites failed at a 9.08% rate. That test has not, as discussed below, been shown to have been validated.
*383But the BES testing procedure is only part of the discrimination practiced against minorities by the JATC. As further illustration, in 1969, in District I there were 95 minority and 142 white applicants. After qualificational procedures, the resulting pool was 47% minority (41 of 88). Among actual selectees, however, minorities constituted 4 of 23. Part of the reason for this result was that the existing waiting list in 1968 was predominantly white.37
It is not merely in gaining access to the JATC’s RA program that a racial disparity appears. Even among the successfully tested and interviewed entrants into the RA program for the years 1973-1974, whites in each respective class worked more hours on the average (168 hours more in 1974 by 1973 white entries; 392 hours more in 1974 by 1974 white entries) and earned more wages on the average ($1,024 more in 1974 by 1973 white entries; $2,331 more in 1974 by 1974 white entries) than minority class members. In 1972, the 1972 white entrants on the average worked 378 hours more than minorities and earned $2,223 more. Dr. Siskin found the pattern of diminished rewards for minorities from these classes to be statistically significant at below the .05 level. This difference in total hours and wages, in context, is corroborative of discrimination even if it is not decisive viewed alone. Furthermore, a larger percentage of minorities than whites have been suspended: 22% versus 6.4% between 1972-1974.38
There has been no evidence of the validity of the JATC tests or the personal interviews. No study of the age and educational requirement has been made by defendants. The educational requirement was historically novel to 542; officials Emrick, Walsh, and Ciavaglia themselves never graduated from high school. No other method of entry requires a high school degree. No creditable evidence has been presented which would show that the educational, age, and interview requirements are a valid basis for making judgments on the fitness of applicants to become operating engineers. Persuasive evidence has been presented indicating the contrary.39 Concerning the validity of the RA testing, we must distinguish the early tests and the JATC test from the BES test. It is only for the BES test that a validation analysis was ever attempted. I find that none of the tests was validly job-related.
Plaintiffs’ expert on this subject was Dr. Richard Barrett, an industrial psychologist with an impressive list of publications and Professor of Management Science at Stephens Institute of Technology. He concluded that the high school educational requirement was not related to job performance. He concluded likewise for the JATC test, since it was not validated, and the personal interview procedure. He concluded that the BES test, which had and was known to have had disparate and discriminatory impact, was not validated by the proffered reports of the United States Training and Employment Service. Barrett determined *384that those reports were based on a sample which was unrepresentative: the average age was forty and experience ranged from 13 months on a given job, for part of this sample, to 178 months total. This is far beyond the limits of experience among RA applicants. Barrett concluded, for these reasons and because of insufficient correlations between test scores and job performance, that the above-referred-to reports did not serve to validate the use of the BES test in this case. I credit his conclusions and find them persuasive. I find further that the reports are not persuasive evidence of validation and hence are insufficient to justify valid use of the BES test.40
In summary, I find the BES test to have contributed to a disparate discriminatory impact in entry to the RA program without a valid job-related purpose. I further find that the statistically significant disparity between minority entries and white entries into the RA program between 1966 and 1972 and the continuing significant disparities between the minority labor pool and the minority RA graduates were the result of overall, non-validated requirements, manipulation of those requirements, and of intentional discrimination participating in the overall pattern of union discrimination previously discussed and proven by statistical and non-statistical evidence.
K. Glasgow, Inc. and the Associations
The ground for liability of defendant class representative Glasgow, Inc. is, in the context of this defendant class suit, principally based upon the legal theory that Glasgow, like other contractors relying on the union’s hiring hall, is liable at least injunctively under 42 U.S.C. § 1981 for the discriminatory acts of the union. This theory also embraces the defendant associations who participated in the creation or approval of an exclusive hiring hall provision. In this sense, the only necessary proof against Glasgow in particular is that it relied contractually upon the hiring hall and is in that respect like other defendant class member contractors working in the 542 jurisdiction. The prerequisite participation of the four named defendant associations has also been proved. This proof, indeed, is not seriously contested.
Plaintiffs have also alleged intentional discrimination broadly throughout the defendant class of contractors and they have similarly alleged a conspiracy among contractors and the union. As discussed below, the intentional discrimination claim and the conspiracy claim are inapplicable in this defendant class context because of the individual questions presented. Plaintiffs have, however, offered much proof pertaining to Glasgow individually and I will discuss this briefly. The conduct of the associations particularly with respect to the “Affirmative Action Program” which substituted for the Philadelphia Plan demonstrated a reckless disregard for equal employment opportunity for minorities but will not be recounted here.
Dr. Siskin studied data pertaining to Glasgow’s employment of operating engineers for the years 1969-1971. He concluded, based upon statistical analyses, that minorities were disproportionately given short-term employment. Although the minority labor pool for Local 542 as a whole is 11.5%, and although much of Glasgow’s work was, as of 1971, centered in District I where the minority labor pool is much higher, between 1969 and 1971 minority individuals averaged only 1.3% of the total hours of employment by operating engineers at Glasgow. For the period 1968-1970, the relative brevity of jobs for minority operating engineer workers is significant below the 1 in ten billion level.
The percentage of minority hirees during any one year from 1969-71, and the corresponding average hours, appears in the following year-by-year table:
*385
To expand upon some of Siskin’s calculations, during the period 1968-1970, whites who were hired in each year worked, on the average, twice as long as minorities hired during the same year. In 1971 minority hirees averaged 190.8 hours while white hirees averaged 334.6 hours.
The number of hirees referred through the hiring hall is not the complete number of Glasgow employees. Some employees serve on a continuing basis year to year. Thus the 1.3% overall rate of minority hours reflects a disproportionately low average number of hours among continuing employees for these years. In 1971 when 9.2% of those hired were minority, only .42% of the hours among continuing employees were minority hours; 80.4% of all the hours worked were worked by continuing, and almost exclusively white, employees. Between 1968 and 1971, Siskin calculated, 117 whites worked continuously more than two years, 22 whites worked more than three years, and only 1 minority worked more than two years. The place of employment of Glasgow’s operating engineer employees is centered in the District I area rather than in one of the more predominantly white districts. In 1971, 70.2% of these employees were assigned there.
Between 1972-1975,41 minorities constituted 4.3% of all those employed, 2.1% of the hours worked and 2.2% of the wages earned. During this period, the percentage of the minority employees reached its highest in 1972 at 4.7%; however, the percentage of hours worked and wages earned by minorities in that year were only 1.2%, the lowest for the four-year period. The average wage among those minorities who were employed was $9.28 per hour as opposed to $8.73 per hour earned by white operators. This slight increase to those employed minorities is insignificant in comparison to the disparities.
Defendants further calculated for the years 1972-1975 reasons for termination among white and minority operating engineers. During the total period, the principal reason for termination was lay off due, basically, to lack of work, weather, or temporary replacements. Whites were terminated by lay-offs at a rate of about 83%, minorities were terminated at a rate of about 80%. Among those discharged for cause 30 were white (1.6%) and only 2 were minority (3.08%).
Apart from the above statistical data, plaintiffs assert instances which they claim reveal discrimination. Plaintiffs also point out that, although Glasgow makes its own selections for master mechanics, none was a minority operating engineer, and hiring by master mechanics were, in the cases of James Swiggard (son of the union’s auditor) and Frank Malloy, discriminatory and impermissibly outside the hiring hall system. Further plaintiffs cite three cases in which minorities were laid off from jobs only to be replaced at those jobs by whites. In another case, that of Willis Fox (see supra), an unidentified master mechanic’s promise of placement on better paying machinery when it became available was disregarded in favor of a white operator. I find that these instances do constitute some proof of discrimination.
It must be pointed out that the studies performed by Siskin were not regression analyses but raw calculations of disparities. This is in contrast to Siskin’s analysis of hours and wages of union members overall. Further, the statistics for 1968-1970 were based on 87.6% of those hired during that period, reflecting the available information.42 It is my conclusion that this “sam*386pie” is not unrepresentative. The 1971 data are based on complete employee information.
Dr. Perl, the expert for the contractors, associations and Glasgow, specifically contradicted the finding of Siskin that the disparities in average wages from 1968-1970 among minority operating engineers were significant at below the 1 in ten billion level. Perl stated the significance was below the .01 level instead. Perl attributed the differences to his rejection of Siskin’s stated standard errors. Perl also undertook a regression analysis for the 1971 data adding a number of factors, particularly skills on specific kinds of machinery. His result was a 126 hour disparity by race which because of the performance of the regression eluded a determination of significance. Plaintiffs’ complain that Perl considered only 156 of the 218 hirees during 1971. Moreover, Perl’s single study of the 1971 data in itself did nothing to rebut Siskin’s analysis for other years.
Although I find in favor of all of Siskin’s above conclusions about Glasgow, the fact remains that he did not attempt a regression analysis which would equalize the factors of age, list, seniority, branch and district. There is thus a possibility that these differentiating factors could account for part of the statistical race effect. Despite the lack of a regression analysis, however, the racially related discrepancies in Glasgow’s workforce and its payroll are patent. They have not been and cannot be satisfactorily explained either with reference to the minority labor pool, the minority representation in Local 542, or any analysis or evidence offered by defendants. The existence of a substantial and almost exclusively white continuing workforce during the years examined is certainly not an excuse for the disparities which such an arrangement fostered in the absence of any persuasive or substantial showing that these workers have characteristics which were job-related and that they are in those respects unlike the available minority labor pool. The suggestion that these men were maintained because they were willing to travel is simply unsupportable.
While substantial evidence has been presented which plaintiffs assert demonstrates intentional discrimination or discrimination in impact by Glasgow, most of this evidence would be more appropriate in the damages stage of this litigation at which time Glasgow’s practices could be evaluated together with the practices of other employers. Thus, I refuse to treat Glasgow separately from the other parties to the contract. Likewise I will not treat the associations separately.
III. Legal Conclusions
Remaining for consideration are the conclusions underlying my decision that the union and the defendant class of contractors and associations are liable at least injunctively for the discrimination practiced through Local 542. Discussed below are the issues pertaining to the certification of plaintiffs’ class, the requirements of Title VII, 42 U.S.C. § 2000e et seq., and 42 U.S.C. § 1981, and finally the matter of the appropriateness of the defendant class.43
A. Appropriateness of Plaintiffs’ Class: Considerations of Standing and the Requirements of Rule 23
The first legal issue which I must consider is whether the named plaintiffs have met the class requirements of rule 23, F.R.Civ.P., and the constitutional Article III requirement of standing. Rule 23(a) defines the standards by which representative parties must be measured.44 To some *387extent these requirements overlap the requirement under Article III that the named plaintiffs be members of the class they seek to represent. At times the courts have appeared to treat adequacy, commonality, typicality and standing almost interchangeably, sometimes treating adequacy as inclusive of commonality and typicality. Defendants’ objections to the plaintiffs’ class and its proposed redefinition touch upon all of these issues. The following discussion will focus first on the issues of adequacy and standing.
1. Adequacy and Standing
Defendants point out that the named plaintiffs were all graduates of the Benjamin Franklin programs and that at the time of certification seven were on the Group III, low priority, out-of-work list while only one had attained union membership and Group I status. The argument is that Group III named plaintiffs’ interests are different from those of members in Group II and Group I. We must look first to East Texas Motor Freight System v. Rodriguez, 431 U.S. 395, 97 S.Ct. 1891, 52 L.Ed.2d 453 (1977), in order to test the certification of plaintiffs’ class.
In East Texas the Supreme Court reaffirmed the view that a named plaintiff cannot, consistent with the standing requirements of Article III and the adequacy requirements of rule 23(a)(4), represent a class of persons if he does not at the time of certification “possess the same interest and suffer the same injury” which he is seeking to prove on behalf of the other class members.45 Id. 97 S.Ct. at 1896; accord, Sosna v. Iowa, 419 U.S. 393, 403, 95 S.Ct. 553, 42 L.Ed.2d 532 (1975). The representative plaintiffs in East Texas were intra-city truck drivers who, because of poor job performance records, simply did not possess the qualifications to become “line drivers.” Plaintiffs in that case alleged discrimination in being barred from transferring to the line driver positions by a no-transfer rule and thus being locked in to jobs which were discriminatorily assigned. The court observed:
[Plaintiffs] could have suffered no injury as a result of the alleged discriminatory practices, and they were, therefore, simply not eligible to represent a class of persons who did allegedly suffer injury. Furthermore, each named plaintiff stipulated that he had not been discriminated against with respect to his initial hire. In the light of that stipulation they were hardly in a position to mount a classwide attack on the no-transfer rule and seniority system on the ground that these practices perpetuated past discrimination and locked minorities into the less desirable jobs to which they had been discriminatorily assigned. [Id. 97 S.Ct. at 1897 (emphasis added)]
East Texas is particularly important to this case because of the qualificational differences among the named plaintiffs here. A member of 542 may not have suffered precisely the same alleged injury of entry discrimination or discriminatory referrals which Group III registrants may have suffered.
But East Texas is distinguishable from the instant case. The representatives in this case are alleging general discrimination in employment practices, principally in membership, application, entry, hours and wages, and referral. Equally important, no class representative has here stipulated that *388there was no discrimination in obtaining his initial status with the union and its hiring hall, and none of the instant representatives suffers from the patently poor job performance record which would on its face preclude advancement to union membership and regular job referral. Potentially, therefore, every class representative and every class member, regardless of his Group position or the status of his union affiliation, has suffered from the alleged single pattern and practice of discrimination and can benefit, together with all others, from the systematic removal of and compensation for discrimination. Thus, the prerequisites for standing are satisfied: the named plaintiffs are members of the class they seek to represent and have alleged the same basic personal injury as that which they allege was suffered by the class. We are also counselled in this conclusion by Hackett v. McGuire Bros., 445 F.2d 442, 446-47 (3d Cir. 1971), which requires in the interest of public policy that standing be interpreted as liberally as Article III permits in Title VII and § 1981 actions.
In addition to the considerations raised by East Texas, in determining whether the named plaintiffs are adequate representatives under rule 23(a)(4), we must also consider the two criteria set forth in Wetzel v. Liberty Mutual Insurance Co., 508 F.2d 239, 247 (3d Cir. 1975):
(a) the plaintiff’s attorney must be qualified, experienced, and generally able to conduct the proposed litigation, and (b) the plaintiff must not have interests antagonistic to those of the class.
By the record in this case on the liability phase it is clear that the plaintiffs’ counsel are qualified, experienced and clearly able to conduct this aspect of the litigation. But this finding should not influence a subsequent trial judge in deciding whether at the damage phase plaintiffs’ counsel can represent all litigants.
The argument that the named plaintiffs are in antagonistic positions is grounded essentially on the view that listees in Group I would profit by minimizing competition from Group III and perhaps Group II. According to this view the advancement of some class members from low priority positions will dilute the interests of those minority persons who are in a priority position. This competition (hence inadequacy) theory, apart from painting a rather grim picture of human nature, is of its own nature highly speculative and could, if adopted generally, defeat much of the effectiveness of the class action device. Nearly every class representative in any context involving disputed facts could in the same speculative way possess the potential to place his own cause above that of the class. In any event, in the present circumstances at the liability stage of the trial the competition theory misperceives the real thrust of rule 23(a) which requires more realistically that there be no “antagonism [arising] . as to the subject matter of the suit.” Redmond v. Commerce Trust Co., 144 F.2d 140, 151 (8th Cir.), cert, denied, 323 U.S. 776, 65 S.Ct. 187, 89 L.Ed. 620 (1944) (emphasis added); accord, Gates v. Dalton, 67 F.R.D. 621, 630 (E.D.N.Y.1975); see 3B J. Moore, Federal Practice ¶ 23.07(3) (1976). As to the issues raised in the complaint and the legal positions which have been or could be taken by any representative at the liability stage there is no conflict. No plaintiff is attacking the concept of a multi-tier experience-graded referral system; rather, plaintiffs attack the failure of the union hiring hall to provide for minorities equal access to and use of that system. Such a claim is consistent with all the representatives’ claims and adverse to none.
Questions of commonality and typicality, which overlap somewhat the question of adequacy, have also been asserted. The argument that commonality has not been shown is based upon the view that, quite apart from the question of conflicts among class representatives, the claims of Group I, II and III named plaintiffs are essentially different from each other and hence not subject to class action treatment together. This analysis is unacceptable. The issue whether the defendants have undertaken a pattern and practice of discrimination is itself essentially a single issue. The prac*389tices alleged against the defendants may have varied somewhat and hence may require separate avenues of proof, some of which will be beyond the familiarity of some of the named plaintiffs, but none of which will be beyond the familiarity of all. The presence of representatives from the various work lists (some possessing attributes of union membership and non-membership, high priority or low priority listing) does not render inconsistent each representative’s claims, but rather assures the kind of diversity necessary to the presentation of all possible avenues of the pattern of alleged employment discrimination. It is in this sense that the plaintiffs’ class is a single class, able to be represented as a whole by the same counsel. The standard in rule 23(a)(2) is therefore duly met.
In much the same sense, the prerequisites of rule 23(a)(3), that the “claims or defenses of the representative parties are typical of the claims or defenses of the class,” are easily satisfied here. The typicality requirement has been interpreted to mean that the “interest of the representative party be coextensive with the interest of the entire class . . . .” Eisen v. Carlisle and Jacquelin, 391 F.2d 555, 562-63 (2d Cir. 1968). Because upon examining the requirement of adequacy and commonality it has appeared clearly that the representatives in this case reflect an interest coextensive with all other representatives and with all members of the entire class and occupy diverse positions in the hiring hall system, there is no reasonable likelihood that legal claims of members of the class or elements of proof at trial would have been overlooked.
2. Rule 23(b)
As the second basic prerequisite to class action certification, some provision of rule 23(b) must also be fulfilled. Inasmuch as the present case involves a plea (central to plaintiffs’ case) for injunctive relief to correct the causes of discrimination in union entry and in the hiring hall process and the discriminatory lack of minority participation, the terms of 23(b)(2) are most clearly applicable:
(b) an action may be maintained as a class action if the prerequisites of subdivision (a) are satisfied, and in addition:
(2) The party opposing the class has acted or refused to act on grounds generally applicable to the class thereby making appropriate final injunctive relief or corresponding declaratory relief with respect to the class as a whole . . . [Emphasis added]
Plaintiffs’ allegation and proof that defendants have engaged in a pattern and practice of racial discrimination necessitates the conclusion that those defendants have allegedly discriminated against the class of minority individuals. The Notes of the Advisory Committee plainly indicate, by way of illustration, that a class-based discrimination claim will typically fall under rule 23(b)(2) so long as damages relief is not the exclusive or predominant relief sought. See 3B J. Moore, Federal Practice ¶ 23.40 at 23-651 (1976). Moreover, the Third Circuit, in Wetzel, 508 F.2d 239, has held that a claim for back pay, as is raised in this case, does not make (b)(2) certification inappropriate where injunctive relief from employment discrimination is also requested.
In addition to certification under (b)(2), certification is appropriate under (b)(1)(A), permitting a certified class where prosecution of separate actions would risk
inconsistent or varying adjudications with respect to individual members of the class which would establish incompatible standards of conduct for the party opposing the class
If individual members of the plaintiff class were left to prove their case against defendants there would not only be gross inefficiency arising from the multiple actions and duplication of proof, but there would be a risk that some, but not all, suing members of the class alleging personal discrimination through systemwide discriminatory policies or practices would be granted injunctive *390relief. This, in turn, could produce the possibility that the defendants as a class would be obliged to provide make-whole relief (perhaps assuring some measure of hours and wages) and correct errant systemwide policies and practices (as by modifications in referral procedure) as to some minority operating engineers while at the same time bearing either no obligation at all or possibly different obligations to other minority operating engineers. This is not a feasible result in the reality of a uniform hiring hall system, the concept of which has not been assailed.
Having given specific attention to the requirements of rule 23, it seems useful now to provide an overview for purposes of perspective. The case of Gibson v. Local 40, Supercargoes and Checkers of the International Longshoremens' and Warehousemens’ Union, 543 F.2d 1259 (9th Cir. 1976), although specifically considering only the 23(a)(3) (typicality) and 23(b)(2) standards, serves as a helpful foil to the range of issues considered in the certification decision in this case. Gibson involved a claim by four clerical workers against an employer association and the union which operated an exclusive hiring hall. The plaintiffs charged discrimination against blacks in the selection of persons to perform clerical duties during the unloading of ships. Clerks were classed into A, B, and casual clerk divisions. Class A, the only class of union members, received priority referrals; then followed the B clerks; and then casual clerks. Plaintiffs were in the lowest category. Although plaintiffs sought to represent all three levels, the Court of Appeals limited the scope of the class to those “black persons who are or may be employed, or who may have attempted or may attempt to obtain employment, as casual clerks . Id. at 1264 (emphasis added).
In reaching its conclusion the court was fully aware that
[a] class action may be maintained . alleging a general course of racial discrimination by an employer or union, though the discrimination may have been manifested in a variety of practices affecting different members of the class in different ways and at different times, [to]
Nevertheless, the court narrowed the class to black casual clerks for the following reasons:
All of the appellants [plaintiffs] were casuals. All held other employment. None of them sought employment as class A or class B clerks or indicated any interest in doing so. The pretrial and trial record concentrated upon the referral and employment of casuals. It is noticeably less complete with respect to the referral and employment of class A and class B clerks. [Id. at 1265]
Most important for the purposes of our present analysis is the conspicuous absence in this case of the factors which weighed against certification of a class of employees from all levels of the hiring hall in Gibson. First, at the time of certification here at least one of the named plaintiffs was a member of 542, a listee in Group I, II or III, or an applicant to the RA program. Second, the Group III and Group II named plaintiffs seek permanent and regular employment as operating engineers and a fortiori seek to advance to a priority status on the out-of-work lists and hence into Group I. All of the named plaintiffs who are not now members of Local 542 desire such membership. The record pre-trial and at trial was fully developed to address the variety of practices which the plaintiffs alleged to constitute a pattern and practice of discrimination. The decisive contrast with Gibson thus reinforces the certification in this case.
3. Subclassing
Remaining for consideration is the purpose and effect of the certification of plaintiffs’ class by subclass upon a motion pursuant to rule 23(c)(4)(B). That rule provides: “a class may be divided into subclasses and each subclass treated as a class, and the provisions of this rule shall then be construed and applied accordingly.” It is apparent that the purpose of this section was to permit the use of a class action device *391even where two or more adverse or potentially adverse classes of plaintiffs were involved. Once such antagonism is noted each class must separately satisfy sections (a) and (b) of rule 23. There is little doubt that this includes the requirements for adequacy purposes, as stated by Wetzel, 508 F.2d 239, that each class be represented by qualified counsel and also that representatives be aligned to the appropriate subclass category. See Wolfson v. Solomon, 54 F.R.D. 584 (S.D.N.Y.1972).
Although in the present case plaintiffs’ original motion for subclassification was granted for skilled minority operating engineers, partially skilled minority operating engineers and as yet unskilled minority persons who wish or may wish to acquire skills in the operating engineer trade, the division was never intended to accommodate antagonism between class representatives or class members. Indeed, paragraph seven of the plaintiffs’ motion to certify subclasses stated that there was no antagonism among the class members. While the division into subclass groups performed the minimally beneficial function of logically describing the different work qualification levels among representatives and class members generally, for all practical purposes the action proceeded just as if no subclasses had been defined. There is thus no conflict with rule 23(c)(4)(B)’s implicit requirement, through 23(a)(4), that each subclass be represented by separate counsel. Because of the absence of antagonism and the existence of a common certifiable class those class representatives and class members who fall into one of the three above-described experience levels were adequately represented by the same counsel.
4. Motion for Amendment of Plaintiffs’ Class Definition
Having concluded that the original certification of a plaintiff class was appropriate we may now consider the question whether under rule 23(c)(1) plaintiffs’ motion to redefine the plaintiffs’ class can or should be allowed. The standard which must be applied, as articulated in Harriss v. Pan American World Airways, Inc., 74 F.R.D. 24 (1977), is “whether the changed class is within the issues tried by express or implied consent by the parties.” Id. at 37; see EEOC v. Detroit Edison, 515 F.2d 301, 310 (6th Cir. 1975), vacated on other grounds, 431 U.S. 951, 97 S.Ct. 2668, 2669, 53 L.Ed.2d 267 (1977). In order to determine whether the proposed definition is permissible the purposes and scope of the changes must be carefully considered. The proposed redefinition is as follows:
(a) For all claims arising under Title VII of the Civil Rights Act of 1964, as amended, 42 U.S.C. § 2000e, the plaintiff class shall include all minority persons (i. e., blacks, Spanish-Surnameds, Orientals and American Indians) who:
(1) Were members, registrants, apprentices or otherwise affiliated with defendant Local 542 at any time between April 30, 1968, and the date of this decree;
(2) Applied for membership or affiliation with defendant Local 542, or who would have applied for membership or affiliation with Local 542 but were deterred from doing so because of Local 542’s discriminatory practices, at any time between April 30, 1968, and the date of this decree; and
(3) Apply for membership or affiliation with defendant Local 542 from the date of this decree until this Court dismisses jurisdiction over this action.
(b) For all claims arising under the Civil Rights Acts of 1866 and 1871, 42 U.S.C. §§ 1981 and 1985(3), and under 29 U.S.C. § 158, the plaintiff class shall include all minority persons who:
(1) Were members, registrants, apprentices or otherwise affiliated with defendant Local 542 at any time between November 8, 1965, and the date of this decree;
(2) Applied for membership or affiliation with defendant Local 542, or who would have applied for membership or affiliation with Local 542 but were deterred from doing so because of Local 542’s discriminatory practices, at any *392time between November 8, 1965, and the date of this decree; and
(3) Apply for membership or affiliation with defendant Local 542 from the date of this decree until this Court dismisses jurisdiction over this action.
Because the proposed amendment arises after trial of this action, there must be a firm reluctance on the part of this court to accept any part of the recertification proposal which might unfairly expand the scope of the class. Comparing the provisions recited above to the original class certification order, set forth in Part I, it is plain that for the most part there is no substantive change. The class as originally certified was divided into the following subclasses: those minority persons in 542’s jurisdiction possessing at least the skills of journeymen operating engineers, those possessing partial skills, and those who wish or in the future may wish to acquire such skills. Thus, members, registrants, and apprentices were included in the overall class. Unsuccessful applicants for membership or affiliation logically fall into at least one of the three original subclasses, and discouraged non-applicants, though not specifically mentioned, fall into the last of the original subclass descriptions.
Defining the scope of the class to reflect the statute of limitations applicable to particular statutory claims, each of which was denoted in the complaint itself, does not on its face produce a substantive expansion of legal claims, for the statutory foundations of the suit implicitly limit the relief available to the class however it is defined. See Wetzel, 508 F.2d 239, 246.46
The argument has been raised, however, that expanding the purview of the class action to encompass pre-1969, pre-Benjamin Franklin I, claims exceeds the standing of named plaintiffs, none of whom asserts any personal injury before that time. Notably Wetzel, in holding that it was proper to define the class with reference to the administrative filing period, did not consider this point. In that case, two female employees alleged Title VII sex discrimination. The court held that persons who were not in a position to make a charge of sex discrimination with the EEOC within 180 days prior to the charge made by named plaintiffs could not be included in the class. But it does not appear from Wetzel that the named plaintiffs had been employed less than 180 days before their filing an EEOC charge. Had they been employed only 90 days, for instance, the court would have had to face the issue presented here — whether unnamed potential claimants who suffered injury prior to the occurrence of any injury to named plaintiffs, but within the limitation period, could be included in a recovering class.
The emphasis in East Texas Motor, supra, on the precise likeness of named plaintiffs claims to those of the plaintiff class members might appear to make the decision of this issue somewhat troublesome. It is clear, however, that pre-1969 issues have been fully illuminated by plaintiffs during the course of this suit. Those issues were directly relevant to the claims of named plaintiffs. To eliminate from the class all members who have suffered an injury from a common source in a common system as proven by common evidence, but who have not suffered that injury at the same time, would require an overly technical view of standing. This aspect of the present case is unlike Davis v. County of Los Angeles, 566 *393F.2d 1334 (9th Cir. 1977), where plaintiffs who had not taken a suspect 1969 employment screening test were barred on standing grounds from representing those who did, even though named plaintiffs had taken a subsequent test. In that case, there would have been little reason for named plaintiffs to pursue proof of discrimination in a 1969 test. Proof of the disparate impact produced by the 1969 test would be irrelevant to the impact of subsequent tests. Here, however, named plaintiffs have alleged “pattern and practice” in addition to a “disparate impact” theory of discrimination as a result of union activities which predated the specific occurrence of their individual claims.
The only other matter for discussion in the redefinition proposal is the specific inclusion of the group of deterred non-applicants. In EEOC v. Detroit Edison the Sixth Circuit Court of Appeals rejected a post-trial attempt to include deterred non-applicants where the district court’s original certification merely created a class of “all black Edison Employees.” Id. at 310. The Court of Appeals contemned this as inconsistent with “fairness . . . [and] the provisions of Rule 23(c)(1).” The court went on to explain by way of dictum that the inclusion of deterred non-applicants would at least require strict subclass separation in order to avoid possibly antagonistic interests between such a group and actual employees. The court’s primary concern was with the adequacy of representation, given the res judicata impact of a class action decision.
Despite the concerns expressed in Detroit Edison, inclusion of the class of deterred non-applicants seems appropriate in this case where one of the original subclass designations included them.
It is certainly clear that persons deterred from applying for employment because of an employer’s reputation for discriminatory activity, where that reputation derives from actual discrimination, can state a cause of action and obtain relief. International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977).47 At issue in the present situation is whether the named plaintiffs, none of whom was a deterred non-applicant, are able adequately to represent such a group. Because the named plaintiffs have been found adequate to the task of proving a pattern and practice of employment discrimination, they are, I believe, at the same time adequate to prove the existence of all those conditions of discrimination which would deter applications from minority persons. The named plaintiffs in proving a pattern and practice of discrimination were proving all that is necessary to be proved at this stage with respect to non-applicants. This result is not unlike that reached in Wetzel, 508 F.2d 239, where the court determined that a former employee was an adequate representative of a class that, included present employees.
On the question whether an individual has actually been deterred, and thus whether he may receive compensation, the Court in Teamsters has indicated that such an individual will himself bear the burden of producing evidence that he is qualified and that he would have applied to an employer but for its reputation for discrimination. Those issues, claims or defenses particular to a deterred non-applicant are not in any event to be considered at this liability stage.
Thus, where the range of issues necessary or appropriate to the establishment of defendants’ liability to deterred non-applicants would be fully and adequately treated by a class of applicants and *394employees in the natural course of the presentation of issues necessary to their own claims and where the class of non-applicants was implicitly within the scope of the original certified class, the inclusion specifically of deterred non-applicants is appropriate. There is no unfairness in extending res judicata effects to the non-applicant group. Doing so does not, moreover, in any way affect the scope of defendants’ proof.
B. Substantive Claims
1. Claims Against the Union
a. Title VII
At the core of this class action is plaintiffs’ Title VII, 42 U.S.C. § 2000e, et seq. (1970), claim against Local 542 for employment discrimination against minority individuals.48 Under the provisions of Title VII a jurisdictional prerequisite to bringing an employment discrimination action in district court is the filing of a charge with the Equal Employment Opportunity Commission, who will then have an opportunity to investigate the charge and take steps to resolve the dispute out of court. As was noted earlier, three black Benjamin Franklin I graduates (Williams, McKay and Muchison), having returned from Resica Falls to be listed in the lowest priority group without experience credit for their six-month training, and having gone two weeks without referrals, filed pro se on June 17, 1969, a charge with the EEOC against Local 542. In this charge they alleged the fundamental facts of their training and complained that they received no referrals due to discrimination based on race. Prior to this charge, in 1968, EEOC Commissioner Vincent Ximines had filed a charge against 542 alleging discrimination against blacks in referrals, the referral system, limitations on and classifications of membership, and entry to the apprenticeship program. The following represents my conclusions with respect to, first, the jurisdictional issues raised by plaintiffs’ Title VII claim, and, second, the merits of that claim.
i. Title VII Jurisdictional Issues
Defendant Local 542, relying upon 42 U.S.C. § 2000e-5(b), argues that in the present case there was no satisfactory attempt at conciliation by the EEOC and that therefore the present action by private individuals is jurisdictionally barred under Title VII. At the time relevant to this case an EEOC charge was required to be filed within ninety (90) days of the alleged discriminatory occurrence. It is not disputed that this jurisdictional requirement was satisfied by the three pro se claimants and Commissioner Ximines. Alternatively defendant asserts that the scope of plaintiffs’ suit must be limited by the allegations of the pro se charge, rather than by Ximines’ charge or by any combination of the two. The conciliation issue will be addressed first.
(a) Conciliation
Under § 2000e-5(b) the EEOC is given a duty to conciliate if it determines that the charge has merit:
*395If the Commission determines that there is reasonable cause to believe that the charge is true, the Commission shall endeavor to eliminate any such alleged unlawful employment practice by informal methods of conference, conciliation, and persuasion.
The Act also provides, pursuant to a 1972 amendment not operative at the time the plaintiffs’ action was filed, that if conciliation has not been reached in 180 days the Commission may itself bring suit. Prior to the 1972 changes in Title VII the Commission had no power to sue.
There have been cases since the amendment in which the absence of a conciliation effort was viewed as a jurisdictional bar. E. g., EEOC v. Hickey-Mitchell Co., 507 F.2d 944, 947-948 (8th Cir. 1974); EEOC v. E. I. DuPont de Nemours & Co., 873 F.Supp. 1321, 1333-34 (D.Del.1974). However, these are cases in which the EEOC was itself the plaintiff. Patterson v. American Tobacco Co., 535 F.2d 257 (4th Cir. 1976), cert, denied, 429 U.S. 920, 97 S.Ct. 314, 50 L.Ed.2d 286 (1976), addresses the issue most concisely:
The 1972 amendments to Title VII empowered the commission to sue if it is unable to secure an acceptable conciliation agreement. This provision of the Act has been construed to create an express condition on the commission’s power to sue.
Id. at 272. The court distinguished EEOC suits from suits by individual employees on their own behalf:
We have held that the commission’s failure to attempt conciliation is not a jurisdictional bar to an employee’s action, because the employee cannot be charged with the commission’s failure to execute its statutory duties.
Id. at 272, n. 14. Accord, Johnson v. Nekooska-Edwards Paper Co., 558 F.2d 841 (8th Cir. 1977); cf. Fekete v. United States Steel Corp., 424 F.2d 331, 335 (3d Cir. 1970). The solution offered by Patterson seems to be the only sensible one. Placing upon individual plaintiffs the burden of pressuring the EEOC to perform its obligations would turn a provision designed to achieve early out-of-court settlement of disputes into a provision perpetuating delay and procedural entanglements to the particular disadvantage of those claiming discrimination. See Occidental Life Insurance Co. of California v. EEOC, 432 U.S. 355, 97 S.Ct. 2447, 53 L.Ed.2d 402 (1977). There is, therefore, no jurisdictional impediment to suit even if the EEOC failed to attempt conciliation.
(b) Scope of Suit
The closer question is whether the 1969 pro se charge or the 1968 Commissioner’s charge or both should serve as the foundation defining the permissible scope of plaintiffs’ § 2000e-2 case. This question is significant because the scope of the charge made to the EEOC may serve to limit the scope of any subsequent suit. See Danner v. Phillips Petroleum Co., 447 F.2d 159, 162 (5th Cir. 1971).
It might be a simple matter to disregard Commissioner Ximines’ charge were it not for the deliberate connection drawn by the EEOC between that and the pro se charge. In its decision letter on the pro se charge the Commission specifically stated that the pro se claimants were as a matter of procedural right (29 C.F.R. § 1601.25b (June 18, 1970)) entitled to receive the Ximines decision. The Commission stated that the pro se claimants should receive that decision because they were aggrieved personally by the discrimination alleged by Ximines and because they were members of the aggrieved class. The fact that James Nunes investigated both cases for the EEOC strengthens the connection. The pro se litigants may thus have been entitled to receive right to sue letters stemming from Ximines’ charge in these circumstances. There is therefore a substantial argument that Xi-mines’ undeniably comprehensive charge should serve to define the breadth of plaintiffs’ suit. But this conclusion is not essential to the decision here.
The only reasonable basis for limiting the scope of a Title VII suit by the charges before the EEOC is to accommodate the conciliation capabilities of the *396EEOC and to prevent the circumvention of the administrative filing requirement. In McBride v. Delta Air Lines, Inc., 551 F.2d 113 (6th Cir. 1977), the court held that “[b]ecause administrative complaints are filed by completing a form designed to elicit specificity in charges, and because the forms are not legal pleadings and are rarely filed with the advice of legal counsel” the subject matter of a district court suit should be limited only to “the ambit of the EEOC investigation that the individual’s complaint might reasonably have been expected to stimulate.” Id. at 115 (emphasis in original); accord, Tipier v. E. I. DuPont de Nemours & Co., 443 F.2d 125, 131 (6th Cir. 1971). It has also been agreed by the Seventh Circuit Court of Appeals that
[t]he correct rule to follow in construing EEOC charges for purposes of delineating the proper scope of a subsequent judicial inquiry is that “the complaint in the civil action . . . may properly encompass any . . . discrimination like or reasonably related to the allegations of the charge and growing out of such allegations.”
Jenkins v. Blue Cross Mutual Hospital Insurance, Inc., 538 F.2d 164 (7th Cir. 1976) (en banc)49 (quoting Danner v. Phillips Petroleum Co., 447 F.2d 157, 162 (5th Cir. 1971)).
Both above-quoted statements of the rule liberally favor pro se claimants. Cf. Haines v. Kerner, 404 U.S. 519, 92 S.Ct. 594, 30 L.Ed.2d 652 (1972). But see Dupree v. Hertz Corp., 419 F.Supp. 764, 769 (E.D.Pa. 1976) (although EEOC charges alleged discrimination in hiring, subsequent court claims alleging discrimination in promotion, assignments, wages and other emoluments of employment were dismissed). The Third Circuit Court of Appeals has also expressed the view that, for purposes of defining the scope of the district court action, technical precision is not a requirement. Wetzel v. Liberty Mutual Insurance Co., 511 F.2d 199, 202-03 (3d Cir. 1975), vacated on other grounds, 424 U.S. 737, 96 S.Ct. 1202, 47 L.Ed.2d 435 (1976). In Wetzel plaintiffs failed to mark a box labeled “Benefits” on the charge form although part of the district court action involved sex discrimination relating to pregnancy benefits. The court took the opportunity to observe:
The private litigant plays an important role in the enforcement of Title VII. The EEOC was created by Congress to effectuate the goals of the Civil Rights Act of 1964 (“Act”). In 1972 the Equal Employment Opportunity Act, 42 U.S.C. § 2000e-5, provided the Commission with further authority to bring its own actions. However, the Act did not provide the Commission with direct powers of enforcement. Since enforcement lies exclusively in the federal courts, Choate v. Caterpillar Tractor Co., 402 F.2d 357, 359 (C.A.7, 1968), the rights of a private party, therefore, must not be barred by procedural technicalities if Title VII is to operate effectively. Alexander v. Gardner-Denver Co., 415 U.S. 36, 44, 94 S.Ct. 1011, 39 L.Ed.2d 147 (1974). Courts have continuously construed Title VII so as not to allow procedural technicalities to bar a claim under the Act. Id.
Guided by the above authorities, the appropriate conclusion is that even without the Ximines charge the pro se charge is sufficient to raise the panoply of mechanisms of the alleged employment discrimination in the present case. The allegation of referral discrimination coupled with the claim that union promises to black Benjamin Franklin trainees were broken do make clear by their very nature that systematic discrimination was being raised. Thus, it is wholly reasonable to expect that *397the resultant EEOC investigation would be systemwide in its attention to employment practices just as it was held in McBride, supra, to be reasonable to expect that a charge of a discriminatory discharge would require investigative review of any racially discriminatory practices.50 In addition, because Commissioner Ximines’ charge was pending (the union having been duly notified of it) and because of the connections between that charge and the pro se charge, it would not be unreasonable to permit the scope of this suit to be defined according to that earlier charge. Any argument that the defendants might otherwise make that permitting the suit as presently constituted is inequitable is substantially undercut by this circumstance.
In summary, the scope of plaintiffs’ suit, as broad as it is in reaching into all possible areas of employment discrimination including entry into Local 542 and referral practices, is supported by the pro se charges of Williams, McKay and Muchison. This conclusion stems in part from the accepted view that technical constructions of a layman’s charge are wholly inappropriate in the Title VII context and in part from the charge itself, which in alleging racial discrimination in referrals (the life-blood of the hiring hall system) and deceptive representations in connection with BFI naturally raises the question for investigation whether there might be a pattern and practice of discrimination.
The asserted lack of conciliation attempts by the EEOC serves as no bar to the plaintiffs’ district court action. The pro se charging parties, two of whom filed their district court action within the requisite thirty days of receipt of their right to sue letters from the EEOC, cannot be held accountable for breaches by the EEOC of its statutory duties on any fair interpretation of the Act.
ii. The Merits of the Title VII Claim
After considering the procedural and jurisdictional objections made by defendant Local 542 to the plaintiffs’ Title VII claims, it is appropriate to turn to the merits of plaintiffs’ case against Local 542.
In Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971), the Court held invalid under Title VII a high school diploma requirement and testing requirements for certain jobs which had the effect of discriminating against minorities.
The Court of Appeals held that the Company had adopted the diploma and test requirements without any “intention to discriminate against Negro employees.” . [G]ood intent or absence of discriminatory intent does not redeem employment procedures or testing mechanisms that operate as “built-in headwinds” for minority groups and are unrelated to measuring job capability. [Id. at 432, 91 S.Ct. at 854]
In International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977), the Supreme Court distinguished between the Griggs- type “disparate impact” case and a pattern and practice suit based on “disparate treatment.” Id. at 335 n.15, 97 S.Ct. 1843. For the former type of case, the Court confirmed that no proof of discriminatory motive is necessary, but for the latter it viewed proof of such motive as “critical.” *398The Court’s distinction is characterized by its subsequent statement that “[e]ither theory may, of course, be applied to a particular set of facts.”
It is not clearly apparent that the Court’s “disparate impact-disparate treatment” distinction is an absolute one rather than a relative or even nominal one. It would seem somewhat inconsistent to require proof only of discriminatory impact where a policy or procedure is innocently neutral on its face but at the same time make mandatory the frequently more difficult proof of discriminatory bias where, in the absence of any identifiably neutral policy, obvious discriminatory effects can be shown. There must be some point, indeed, at which either theory may be applied. Nevertheless, plaintiffs have argued that they satisfy all the necessary criteria of a pattern and practice suit of intentional discrimination. The present analysis will focus upon both theories.
(a) Intentional Discrimination
Teamsters requires that a Title VII plaintiff seeking to prove intentional discrimination establish by “a preponderance of the evidence that racial discrimination was the company’s standard operating procedure— the regular rather than the unusual practice.” Id. at 336, 97 S.Ct. at 1855. As has been discussed above, plaintiffs’ burden of proving a prima facie case of intentional discrimination may be satisfied in part or in whole by evidence showing a gross statistical disparity between a particular employer’s workforce (or union’s membership) and the racial composition of the relevant community population. See id. at 339-40 n.20, 97 S.Ct. 1843. In Hazelwood School District v. United States, 433 U.S. 299, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977), the Court reaffirmed this view emphasizing in plain terms that for some jobs which require a special qualification the correct statistical comparison is between the racial composition of an employer (or union) and the percentage of minority persons in the labor market consisting of those persons in the community possessing the necessary qualifications. Id. 97 S.Ct. at 2742 n.13.
Plaintiffs have presented before this court a wide variety of statistical evidence assembled by Dr. Bernard Siskin, an experienced expert in the field of employment statistics who has demonstrated his thorough competency and credibility in this matter. As has been explained above, Dr. Siskin conservatively estimated that the minority labor pool of men between the ages of 18 and 65 constitute 11.5% of the total labor pool. Because beginners in the operating engineers trade traditionally have not and logically do not require any special educational background or prior skills, notwithstanding 542’s use of educational and testing requirements in the registrant and JATC programs, there is no need to redefine the labor pool for these or other factors urged by defendants.
In light of all the factual statistical conclusions reached above, plaintiffs have established on a preponderance of the evidence that a gross statistical disparity existed. The disparity between labor pool representation and union membership in 1971 is virtually unexplainable by any theory of random selection; the chance of random occurrence is less than 1 in 1023, far less than that needed to establish statistical significance. This disparity is confirmed and corroborated by other statistical analysis discussed above.
If it were necessary to rely upon these data alone, it may well be that plaintiffs’ prima facie burden of intentional discrimination would be met. But other proof offered by plaintiff and credited by this court confirms in non-statistical ways the strong statistical inference of purposeful discrimination. This proof, detailed above, consists of the union activities leading to the obviation of the Philadelphia Plan, particularly the repeated gross inaccuracies with respect to calculations of the minority representation in the union, the union’s treatment of minority Benjamin Franklin graduates who trained for six months in the reasonable belief that they would at least receive experience credit in the referral system, discretionary departures from legiti*399mate entry methods disproportionately favoring white entries, tendencies to the creation of segregative methods of entry, discretionary and arbitrary referrals and the numerous instances of discriminatory actions against individual minority operating engineers. See Casteneda v. Partida, 430 U.S. 482, 496, 97 S.Ct. 1272, 51 L.Ed.2d 498 (1977). This total fabric of non-statistical proof, together with the statistical disparities, leads to only one conclusion — that plaintiffs’ proof easily met the requirements of a prima facie case. The proof offered in rebuttal has not altered my view of the persuasiveness of plaintiffs’ case.51
(b) Disparate Impact Discrimination
Although Griggs involved facially neutral testing and educational qualifications which had the effect of excluding blacks at a much higher rate than whites, the Court was expressly guided by the view that “Congress directed the thrust of the Act to the consequences of employment practices, not simply the motivation.” 401 U.S. at 432, 91 S.Ct. at 854. The statistical nature of the proof here is somewhat different from Griggs in that there is no single identifiable standard which produces direct and clearly ascertainable disparate rates of membership, wages or hours. But the significance of the statistical disparity as expressed in competent expert testimony makes plain that 542’s practices have produced a seriously disproportionate racial consequence. Because defendants have offered no significant evidence (in the face of this disparity) that any of the practices relating to the selection for affiliation or membership and dissemination of referrals, etc., are job related, an analogy to Griggs indicates that plaintiffs should prevail. See Gibson, supra at 1265 (holding that a claim of discrimination in a hiring hall job referral setting need not require proof of discriminatory intent).
While framing the issue in this way may seem a bit unusual, especially where the evidence of disparate effect may itself be so gross as to create an inference of intentional discrimination, there is no reason to overstate the requirements of proof and thereby perpetuate an anomaly in legal theory by which the evidentiary standards become more stringent where discriminatory effects cannot be explained by an apparently innocent policy or practice. It is my conclusion, therefore, that had the plaintiffs failed to meet the rigorous burden of proving intentional discrimination, a Griggs -type claim was adequately proved.
b. Section 1981 Claims Against Union
In addition to their Title VII claim against 542, plaintiffs have argued that liability flows from 42 U.S.C. § 1981, which provides:
All persons within the jurisdiction of the United States shall have the same right in every State and Territory to make and enforce contracts, to sue, be parties, give evidence, and to the full and equal benefit of all laws and proceedings for the security of persons and property as is enjoyed by white citizens, and shall be subject to like punishment, pains, penalties, taxes, licenses, and exactions of every kind and to no other.
In recent years § 1981 has emerged as an important provision in employment discrim*400ination suits. See, e. g., Young v. International Tel. & Tel. Co., 438 F.2d 757 (3d Cir. 1971). Johnson v. Railway Express Agency, Inc., 421 U.S. 454, 95 S.Ct. 1716, 44 L.Ed.2d 295 (1975), has made clear that § 1981 is not delimited by the procedural prerequisites of Title VII. See Alexander v. Gardner-Denver Co., 415 U.S. 36, 47 & n.7, 94 S.Ct. 1011, 39 L.Ed.2d 147 (1973).
What has remained undecided by the Supreme Court is the appropriateness of specifically analogizing between the standards of proof required under Title VII, including the standard by which only a disparate impact need be shown, and those under § 1981. Although the foregoing finding of intentional discrimination on the part of 542 with respect to plaintiffs’ Title VII claim is decisive under § 1981 as well, it remains to be examined whether § 1981 also permits a finding of discrimination upon proof of disparate impact alone.
The leading case holding that a § 1981 claim requires only proof of discriminatory impact is Davis v. County of Los Angeles, 566 F.2d 1334 (9th Cir. 1977), cert, granted, 437 U.S. 903, 98 S.Ct. 3087, 57 L.Ed.2d 1132 (1978). That case in effect adopted for § 1981 purposes the rationale of Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971), that facially neutral employment practices which produce a disparate racial effect are actionable under Title VII. Both Griggs and Davis v. County of Los Angeles involved employment test procedures and requirements which were not validated as job-related. Specifically, Davis v. County of Los Angeles enjoined the use of a non-validated standardized test and a 5'7" height requirement for fire department employees. Both the test and the height requirement caused a disproportionate exclusion of Mexican-American job applicants.
Davis v. County of Los Angeles is significant in that it comes after Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976), which expressly denied relief from an alleged discriminatory police employment testing procedure in the absence of proof of discriminatory motive. Although Washington did not involve a Title VII claim, the Court’s holding did specify that the denial of relief was based on constitutional (fourteenth amendment equal protection clause) and statutory grounds: “Respondents were entitled to relief on neither constitutional nor statutory grounds.” Id. at 248, 96 S.Ct. at 2052. Justice Stevens, concurring in part, observed that the “actual statutory holdings are limited to 42 U.S.C. § 1981 and § 1-320 of the District of Columbia Code . . . .” Id. at 255, 96 S.Ct. at 2055. An argument does exist, therefore, that Washington may have foreclosed the issue, requiring that discriminatory motive be proven as a prerequisite to § 1981 relief. However, the Ninth Circuit in Davis v. County of Los Angeles distinguished Washington first by citing to apparent limitations in Justice White’s majority opinion:
Mr. Justice White prefaced Part II of the majority opinion with this statement: “Because the Court of Appeals erroneously applied the legal standards applicable in resolving the constitutional issue before it, we reversed . . . .” Id. at 239, 96 S.Ct. at 2046 (emphasis added).
566 F.2d at 1339. The thrust of the Ninth Circuit’s rationale is that Washington never specifically discussed the § 1981 considerations and hence did not decide the scope of § 1981 liability. The court stated: “In the absence of any express pronouncement from the Supreme Court — a pronouncement not delivered in Washington — we are unwilling to deviate from this established practice [of equating Title VII with § 1981].” 566 F.2d at 1340.
It is interesting to note, however, that the dissenting judge on the Ninth Circuit’s panel (Wallace, J.), while agreeing that Washington v. Davis was not decisive as to the § 1981 claim, stated that the legislative history of § 1981 should track the fourteenth amendment’s standards of proof. Section 1981 was, however, first enacted pursuant to the thirteenth amendment before passage of the fourteenth amendment. It was then simply reenacted in 1870 after the fourteenth amendment’s *401passage to reassure that its provisions would not be held unconstitutional. Hence the linkage between the elements of a § 1981 cause of action and those of an action based on the equal protection clause seems unfounded. I therefore adopt the view that a § 1981 employment discrimination claim may be proven on roughly the same basis as a Title VII claim, including proof of disparate impact alone. In this view I agree with the majority in Davis v. County of Los Angeles and add that § 1981’s close connection to the thirteenth amendment amply distinguishes the fourteenth amendment decision in Washington.52
c. Section 1985(3)
Another legal aspect of plaintiffs’ ease focuses upon the alleged conspiracy among the union, associations and contractors to discriminate intentionally against minorities. Because this claim depends largely upon the involvement of the associations and contractors, full discussion will await the consideration of their liability below.
2. Claims Against Associations and Contractors
a. Section 1981
While I have found the union liable under § 1981, a separate adjudication must be made as to whether the union’s conduct in operating the hiring hall has caused the associations and contractors to be liable under that section. The evidence is not clear as to the quantum of specific and personal knowledge which the contractors and their associations may have had of the union’s acts of intentional discrimination or the impact of their hiring hall policies. The fact is that the vast majority of individual contractors never hired a minority operating engineer; that the GBCA, CAEP, PECA and UCA signed a statement, relevant to federal approval of the “Affirmative Action Program” displacing the Philadelphia Plan, grossly exaggerating minority union membership; and that the gross disparity between the percentage of the minority representation in the labor pool and minority representation in the union along with a gross disparity in hours and wages of minorities as against the minority labor pool percentage is a matter of such broad scope that some or all contractors and associations might have had knowledge of it. Yet I find that the plaintiffs have failed to prove on a preponderance of the evidence that the associations or contractors viewed simply as a class were actually aware of the union discrimination affecting the employment of minority persons throughout the operating engineer industry in Local 542’s jurisdiction. Such knowledge with actual participation in the hiring hall system would of course itself constitute intentional discrimination. I also find that not all contractors and associations may be said as a class to have had reasonable notice of the union’s discrimination in view of the great number of contractors and the varying size and intensity of their work. Nevertheless, I find that the contractors and associations are injunctively liable to the plaintiff class under § 1981 as a result of their contractual relationship to and use of a hiring hall system which in practice effectuated intentional discrimination, whether or not the employers and associations knew or should have known.
Although no case has been found which imposes civil rights liability on the employer or association in these precise circumstances, there are several cases in which acts by union hiring halls favoring union members in violation of 29 U.S.C. § 158(a) of the National Labor Relations Act have created liability in an employer who had neither knowledge nor reasonable notice of such/ acts. There are a few district court decisions in civil rights contexts, however, which hold or suggest that no such extension of liability from union to employer may be made. Added to this legal fabric are cases which, in interpreting the scope of Title VII or § 1981 liability, have held em*402ployers liable for the discriminatory impact of facially neutral terms of a collective bargaining agreement governing a hiring hall’s operation.
My conclusion that as to the § 1981 claims the association and contractors can be liable for at least injunctive relief is predicated on my analysis of: (1) the analogous National Labor Relations Act cases; (2) the lower court Title VII and § 1981 decisions involving an employer who had no knowledge or reasonable notice of hiring hall discrimination against minorities; (3) Title VII cases in which an employer’s agreement to a facially neutral hiring hall system had a discriminatory impact; and (4) cases and principles relating to the applicability of responden t superior liability under § 1981.
1. The NLRA Cases
The plaintiffs seek to hold the class of defendant contractors liable vicariously for the acts of the union in its hiring hall operation. The rationale for such a view is that the employers by agreeing to a hiring hall arrangement have made the union their agent in hiring. See NLRB v. H. K. Ferguson Co., 337 F.2d 205, 208 (5th Cir. 1964), cert, denied, 380 U.S. 912, 85 S.Ct. 898, 13 L.Ed.2d 798 (1965); Morrison-Knudsen Co., Inc. v. NLRB, 275 F.2d 914, 917 (2d Cir. 1960) (Swan, J.), cert, denied, 366 U.S. 909, 81 S.Ct. 1082, 6 L.Ed.2d 234 (1961); cf. NLRB v. United States Steel Corp., 278 F.2d 896, 898 (3d Cir. 1960) (en banc), cert, denied, 366 U.S. 908, 81 S.Ct. 1083, 6 L.Ed.2d 234 (1961). I am persuaded by the rationale expressed most clearly in Morrison-Knudsen and hold that that rationale is applicable to the instant § 1981 action.
Morrison-Knudsen involved discrimination by the union hiring hall against nonunion workers. Such a practice was and is forbidden by 29 U.S.C. § 158(a)(1). The court’s view was expressed in the following quotation:
[RJegardless of the extent of their [the employers’] knowledge we agree with the Board that an employer may not avoid liability for violations of the Act by the hiring hall when he has turned over to it the task of supplying the men to be employed. The Local acted as agent for the petitioners in selecting the men to be hired. Its discriminatory acts, which unlawfully encourage membership in Local 545, are properly chargeable to the agent’s principal as discriminatory acts by it. [275 F.2d at 917 (emphasis added)]
See H. K. Ferguson, supra, at 203.53 Although not discussing the basis for its agency conclusion in detail, it seems clear that the Morrison-Knudsen court was concerned with the federal policy implications of immunizing an employer from liability for violations of the NLRA arising out of a function so typical to employers.
Of course, Morrison-Knudsen’s expression of a vicarious liability rule has not been followed by all courts of appeals. NLRB v. Master Stevedores Association of Texas, 418 F.2d 140 (5th Cir. 1969) (Godbold, J.), and Lummus Co. v. NLRB, 119 U.S.App.D.C. 229, 339 F.2d 728 (1964), emphasize that Morrison-Knudsen prefaced its statement of the vicarious liability with the following observation:
While there is some dispute as to the extent of the petitioners’ knowledge of the Local’s discriminatory practices, it seems clear at least they knew of the collection of “dobie” [a license fee for non-union workers]. [Id. at 917]
However, in its footnote 6, the court in Morrison-Knudsen explained the basis for its conclusion that the employers knew of the “dobie”, quoting testimony, apparently by some managerial official, that he told all stewards that no dues, subscriptions, donations, or “dobie” were to be collected on the *403job. Although this evidence might have shown knowledge that some fee was required of non-union workers or workers from other unions by the local operating the hiring hall, the court’s finding of discrimination appears to have rested in part on the proven fact that the “dobie” was higher than the dues charged of union members, and there is no indication that the employer knew of the discriminatory amount of the fees. The court also affirmed the finding that the union operating the hiring hall discriminated by giving direct preference to its union members in referrals. Certainly knowledge of “dobie” is not knowledge of such preference. Thus it seems inescapable that the reason Morrison-Knudsen expressed its vicarious liability foundation was that the knowledge of “dobie” alone simply did not constitute adequate knowledge or notice of the discrimination alleged.54
Master Stevedores, supra, in requiring a knowledge or reasonable notice standard did not attempt a searching analysis of Morrison-Knudsen’s knowledge reference. But in Pacific Maritime Association v. NLRB, 452 F.2d 8, 10 (9th Cir. 1971), the court was unwilling to minimize Morrison-Knudsen’s vicarious liability holding. In that case, involving hiring hall discrimination against non-union workers, the court found that an employer association knew or should have known of the discriminatory union practices.
In a slightly different NLRA context, the Third Circuit in NLRB v. United States Steel Corp., 278 F.2d 896, 898 (3d Cir. 1960 (en banc), cert, denied, 366 U.S. 908, 81 S.Ct. 1083, 6 L.Ed.2d 234 (1961), also held an employer liable for union favoritism via a hiring hall system where a union master mechanic, already in the employ of a contractor, was permitted to select union members to fill vacant positions at the employer’s plant. Although the employer did not sign the bargaining agreement which empowered the master mechanic to make selections, it did acquiesce in following that procedure. The court stated:
We think the Board was justified in concluding that, as a result of the Master Mechanic’s dual capacity as agent for both American Bridge and Local 542, both employer and union were responsible for the hiring procedures employed. [Id. (emphasis added).]
United States Steel is a case in which the court reached the conclusion that a union’s hiring method could create liability in the employer on a theory of agency.55
In summary, I find the Morrison-Knudsen rationale to be applicable here. The decision in United States Steel that a master mechanic was agent for both the employer and the union further indicates the appropriateness of finding the hiring hall an agent of the employer in an NLRA setting. I see no reason — statutory or policy — why this conclusion should be any different in a civil rights setting.
2. Civil Rights Cases Rejecting Vicarious Employer Liability
In contrast to the rationale of Morrison-Knudsen, defendants rely on certain § 1981 and Title VII cases which purportedly hold that there can be no automatic employer liability upon proof of a union’s discriminatory operation of a hiring hall. Allen v. Pipefitters Local Union No. 208, 56 F.R.D. *404473 (D.Colo.1972) and Vogler v. McCarty, Inc., 294 F.Supp. 368 (E.D.La.1968) aff’d on other grounds sub nom. Local 53 v. Vogel, 407 F.2d 1047 (5th Cir. 1969). While I recognize that these cases are partially in conflict with my holding that employers are liable for the acts of a hiring hall, I find them to be unpersuasive and, in some respects, factually and analytically distinguishable.
In Allen, a § 1981 and Title VII case, the court held that two employers’ associations were not rendered liable by entering into a collective bargaining agreement with a union where the agreement provided that the union would operate an exclusive hiring hall and where the alleged discriminatory acts by the union did not arise out of the terms of the agreement. The acts of discrimination alleged against the union related to internal membership policies and consisted of the following: limitations on the number of new members; use of non-validated written tests; a requirement that an entrant be recommended by a union member; a subjective evaluation by an Executive Board; and a requirement that entry be approved by an affirmative voice vote of the membership. The court concluded:
The essence of the allegations as to these associations is that they are parties to a collective bargaining agreement executed with Local 208 and hire virtually all of their employees by referral from Local 208’s Hiring Hall.
These allegations are not sufficient to state a claim against the associations . . . . [56 F.R.D. at 474-75]
Notably, the court cited to no authority in reaching this result.
In Vogler, the court in similar circumstances stated that two employers could not be held liable under Title VII for the discriminatory membership policies of the union which, under the collective bargaining agreement, operated an exclusive hiring hall for the employers. The court expressed its view of the positions of the employers:
Local 53’s operation of a hiring hall arrangement which, together with its membership admission policies, have controlled employment within the insulation industry in the New Orleans and Baton Rouge areas and has resulted in discrimination against Negroes and other minority groups seeking employment through the Union’s hiring hall does not result in a violation of Title VII . by employers who are required to accept for employment only those union members or persons to whom the union issues work permit cards. [294 F.Supp. at 374-75]
The court simply did not believe that the hiring hall agreement in any way created a responsibility in the employers for its execution. As in Allen, the court here cites no authority for this legal conclusion.
Vogler is an extremely difficult case to analyze in detail because the court’s opinion seems incomplete. It does appear clearly, however, that three private claimants (Vogler, Galaviz, and Joseph) sought to join two employer companies (McCarty and Branton) to a suit by the United States under Title VII for a preliminary injunction against a union based on hiring hall discrimination. Most of the claims in Vogler were dismissed or dismissable for lack of jurisdiction. With this in mind, and given the lack of citational authority on the non-jurisdictional dismissal, Vogler in and of itself is unpersuasive.
Three other cases also reach conclusions which non-parties have construed as consistent with the holdings in Allen and Vogler. These cases are: Rios v. Enterprise Association Steamfitters, 400 F.Supp. 988 (S.D.N.Y.1975), aff’d in part rev’d in part sub nom. EEOC v. Enterprise Association Steamfitters, 542 F.2d 579 (2d Cir. 1976), cert, denied, 430 U.S. 911, 97 S.Ct. 1186, 51 L.Ed.2d 588 (1977); Guerra v. Manchester Terminal Corp., 498 F.2d 641 (5th Cir. 1974); and Byrd v. Local Union No. 24, 375 F.Supp. 545 (D.Md.1974).
In Rios the court was presented with a motion for back pay after a “protracted litigation under Title VII.” 400 F.Supp. at 990. Plaintiffs sought an award of damages from a contractor’s trade association and the court denied the request:
*405MCA, a trade association of certain contractors in the New York area, acts only in collective bargaining negotiations between its members and Local 638. MCA does not employ steamfitters; rather, employment is done by its members [without a hiring hall]. While MCA was found to have been properly made a party defendant in the Rios action . [United States v. Local 638, 360 F.Supp. 979, 994-95 (S.D.N.Y.1973) (Bonsai, J.)] that finding did not imply that MCA was “responsible ipso facto for all the employment . . . practices here found unlawfully discriminatory or . liable in damages to the plaintiffs in Rios. Plaintiffs have shown no specific instances of MCA discrimination . . . .”
Id. at 992. The court, in the earlier opinion (360 F.Supp. 979) quoted from above, did include the association as a Title VII defendant and did include it in its order for future action to correct a lack of non-white employment in the building trades industry.56 360 F.Supp. at 996. The court pointed out in that same opinion:
MCA has greater influence over and responsibility for employment practices applying to the industry as a whole than any single employer. Moreover, the participation of MCA in an affirmative action program is a necessity if the steamfitting industry is to correct the discriminatory effects of past employment practices.
Id. at 995.
The district court decision in Rios can only truly be said to relate to the inappropriateness of a back pay award against an association where that association has not itself been proven to have actively discriminated. Inferentially, this means that in Rios the discriminatory acts of the union (not found to be intentional) would not be imputed for purposes of back pay to an association of contractors which acted primarily in negotiating bargaining agreements and which committed no specific acts of discrimination of its own.
But in affirming this aspect of Judge Bonsai’s decision, the Second Circuit Court of Appeals in EEOC v. Enterprise Association Steamfitters, supra, was circumspect on the issue of absolving the MCA from back pay liability. The plaintiffs argued that, by enjoining MCA, the district court must have found a sufficient basis for holding the association to account for discrimination; that same basis, they argued, must perforce support an award of back pay. The Court of Appeals added:
The Rios plaintiffs argue, not without reason, that this [injunctive] relief could not have been ordered absent a finding of discrimination by MCA, purposeful or otherwise, and that all that is required to establish MCA’s liability for backpay is the finding of discrimination; indeed, mere acquiescence in the discriminatory acts of the union would render it liable. [Citations omitted] Absent any specific finding of discrimination, however, we conclude that the district court’s finding of nonliability on the part of MCA is not an abuse of discretion. Guerra v. Manchester Terminal Corp., 498 F.2d at 655-56. [542 F.2d at 589 (emphasis added)]
It may be inferred from the above dictum that acquiescence (presumably involving some meaningful level of knowledge or notice) constitutes sufficient involvement by a contractor’s association to require holding it liable for back pay relief under Title VII. But it may be inferred from the holding that either back pay liability or injunctive liability without back pay may be imposed in the absence of “acquiescence” and without specific discriminatory acts in the discretion of the trial court. Certainly neither the district court opinion nor the Court of Appeals’ opinion in Enterprise Association Steamfitters supports the conclusion that an association or employer bearing a legal *406contractual relation to a union is insulated from liability for the union’s discriminatory acts in the absence of participation in or knowledge of such acts. Indeed, the Court of Appeals’ affirmance of injunctive liability in the absence of “any specific finding of discrimination” supports an opposite conclusion.
The decision in Guerra, supra, is likewise distinguishable from Vogler and Allen. Guerra was a case in which a union was held exclusively liable under § 1981 for backpay because of discriminatory hiring hall practices against aliens. In that case the employer entered into a collective bargaining agreement giving hiring hall preference to United States citizens. Both the union and the employer were found to have violated § 1981. The district court, however, in its discretion imposed damages liability on the union alone because of its view that the union was principally at fault. Thus, Guerra supports the view that employers’ monetary liability for hiring hall violations may be subordinated to the union’s liability even where the terms of the collective bargaining agreement itself are discriminatory in purpose. But Guerra does not as a matter of law absolve employers in all such circumstances, nor does it speak to the issue whether an employer may ever be held liable injunctively or monetarily for discriminatory union acts not rising out of a bargaining agreement’s terms.
The case of Byrd v. Local Union No. 24, supra, also falls short of supporting the holdings in Allen and Vogler that an employers’ association, and inferentially an employer, cannot be held liable absent knowledge or notice of a union’s discriminatory acts in operating a hiring hall under a lawful collective bargaining agreement. The court first turned to an officially unpublished bench opinion in McFadden v. Baltimore STA, Civ. No. 71-457-H, bench opinion (D.Md. July 7, 1972) (8 F.E.P. 391), which apparently held that a trade association could not be held liable for union discrimination where the association merely agreed to a union hiring hall. In Byrd the plaintiff alleged
that the respective contractor associations have knowledge of the alleged actions taken by their agents, the JATCs, which discriminate against the plaintiffs and their putative class on account of their race and that the alleged discriminatory acts [footnote omitted] of their respective agents, the JATCs, have been ratified and encouraged by the contractor associations. ... A final allegation is that the defendant contractor associations have engaged in discriminatory collective bargaining agreements with the . unions . . . . [375 F.Supp. at 560-61]
The court, considering the case on a motion to dismiss, distinguished McFadden and Allen saying “the alleged liability of the contractor associations is based upon their own respective acts vicariously performed by their alleged agents . . . .” Id. at 562. The importance of Byrd as possible support for the Allen and Vogler position (that an employer or association of employers is not liable vicariously for discrimination by the union in a hiring hall) is diminished because of the allegations in Byrd of direct acts by the employer association.
In summary, Vogler and Allen, although directly rejecting a vicarious liability theory in a civil rights context, do not represent a decisive analysis of the issues and cite no authority for their holdings. Byrd, while in apparent agreement with the Allen and Vogler position, did not have to reach the issue of vicarious liability because of allegations of direct acts by the employer associations.
In Rios, a slightly different type of case, the district court’s decision was to absolve an association of monetary (but not injunctive) liability where no act by the association was shown to be connected to the union’s hiring discrimination. On appeal, that aspect of Rios was affirmed with the observation that acquiescence by an association would provide a sufficient basis for back pay liability but that because no specific act of discrimination was proven against the agency the refusal by the district court to hold it liable for back pay was not an abuse of discretion. Similarly, Guerra was a case *407in which the Court of Appeals declined to view as an abuse of discretion the district court’s refusal to require an employer to pay monetary damages under § 1981 where, although both employer and union agreed to a hiring hall system which was itself discriminatory, the union was principally responsible.
In view of this reading of the civil rights cases involving a union hiring hall, there is no persuasive reason to reject the theory of vicarious liability of employers for discrimination by the union in its hiring hall operation. Indeed the Second Circuit Court of Appeals decision in Enterprise Association Steamfitters supports the vicarious liability theory.
3. Cases In Which an Employer or Employer’s Association Was Held Liable for Discrimination Arising From Terms of Contract
There do exist cases in which, either under Title VII or § 1981, participation in a collective bargaining agreement having a discriminatory impact justified imposition of liability upon an employer, an employer association or a union, even though there was no specific showing of an intent to discriminate. E. g., Myers v. Gilman Paper Corp., 544 F.2d 837 (5th Cir.) cert, dismissed, 434 U.S. 801, 98 S.Ct. 28, 54 L.Ed.2d 59 (1977); Carey v. Greyhound Bus Co., Inc., 500 F.2d 1372 (5th Cir. 1974); Johnson v. Goodyear Tire & Rubber Co., 491 F.2d 1364 (5th Cir. 1974); Robinson v. Lorillard, 444 F.2d 791 (4th Cir.), cert, dismissed, 404 U.S. 1006, 92 S.Ct. 573, 30 L.Ed.2d 655 (1971); National Organization for Women, Inc. v. Minnesota Mining and Manufacturing Co., 73 F.R.D. 467 (D.Minn.1977). Each of these cases, however, involved seniority systems arising under the collective bargaining agreement and in each of the cases it was the seniority system which was said to have perpetuated past discrimination. Importantly, these cases were decided prior to the Supreme Court’s decision in International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977), which upheld as lawful seniority systems of a bona fide character though they may in effect perpetuate preAct discrimination.57
Nevertheless, the above-cited cases still have important precedential impact. Apart from the narrow question whether a bona fide seniority system provided for by contract is statutorily exempt from liability, these cases instruct that all parties to a collective bargaining agreement which by its terms has a discriminatory impact can bear civil rights responsibility.
It must be emphasized that these decisions do not explicitly resolve the question whether under § 1981 a union hiring hall’s acts may be imputed to an employer for liability purposes in the absence of proof of knowledge, notice or intent. The eases are decided either under Title VII, which according to Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971), creates liability merely from discriminatory effects of facially neutral hiring practices, or under § 1981 with the view (as accepted here) that liability requires no proof of purposeful conduct on the part of any of the defendants. These cases give analytical support to plaintiffs’ vicarious liability theory in the present case.
In Myers for instance, a case in which the court found that a facially neutral seniority system provided for in the collective bargaining agreement had a discriminatory impact on transfers, 544 F.2d at 848, the court affirmed the district court’s holding that the union and the employer should share backpay liability while reserving for the completion of a Stage II proceeding the question of the exact apportionment of *408backpay between them. The Myers court also held an international union liable under Title VII and § 1981 because one of its representatives attended negotiations between the employer and the local during one contract session and because the resulting contract was (as was the previous one) approved by the international’s president. The court, referring to the liability of the international, observed:
Labor organizations, as well as employers, have an affirmative duty to take corrective steps to prevent the perpetuation of past discrimination. [Citations omitted.] “It is the reasonably certain prospect of a backpay award that ‘provide^] the spur or catalyst which causes employers and unions to self-examine and to self-evaluate their employment practices and to endeavor to eliminate, so far as possible, the last vestiges of an unfortunate and ignominious page in this country’s history.’ ” Albemarle Paper Co. v. Moody, 422 U.S. 405, 417-18, 95 S.Ct. 2362, 2373, 45 L.Ed.2d 280 .. . (1975). [Id. at 850-51]
The court’s reference to an affirmative duty is most significant. It would appear that even where the party to a collective bargaining agreement foresees and intends no discrimination, if a term in that agreement discriminates in impact, he is still to be held liable. In other words, the duty described in Myers is a duty to be aware of potential discriminatory impact resulting from an agreement which does not on its face reflect any discrimination at all.
Carey v. Greyhound Bus Co., Inc., supra, also involved a facially neutral but in effect discriminatory seniority provision. Once again holding all parties liable, the court stated:
[U]nion contracts grant no immunity on the subject of racial discrimination. Neither can the employer use the union or unions for a shield. . . . Writing in Peters v. Missouri-Pacific R.R. Co., [483 F.2d 490 (5th Cir. 1973)] Judge Godbold said:
“Even if the railroad’s intent in agreeing to and enforcing the challenged employment agreement was a racially neutral, or even benevolent, desire to discharge its legal duty to bargain with the plaintiffs’ representative and to enforce the contract entered into, such intent does not save the agreement from the terms of Title VII if it operates to ‘freeze’ the status quo of prior discriminatory employment practices.” [500 F.2d at 1377]
The policy expressed in Carey, a case which was one of the building blocks for Myers, suggests an absolute liability theory.
Robinson v. Lorillard, a case also involving a facially neutral seniority system, seems consistent with this interpretation:
The rights assured by Title VII are not rights which can be bargained away — either by a union, by an employer, or by both acting in concert. Title VII requires that both union and employer represent and protect the best interests of minority employees. Despite the fact that a strike over a contract provision may impose economic costs, [footnote omitted] if a discriminatory contract provision is acceded to the bargainee as well as the bargainor will be held liable. [444 F.2d at 799]58
Acknowledging that in each of the above cases the union and employer are interconnected via the bargaining agreement, which itself is the cause of the discriminatory impact in violation of § 1981 or Title VII, these decisions make an important point. If the state of mind of the parties, being presumed neutral, cannot establish a defense to a Title VII or § 1981 violation, it is *409a short conceptual step to hold liable an employer who agrees to a facially neutral agreement with a union who proceeds to create a discriminatory impact by acting (intentionally or not) outside the agreement. It may be said that an employer’s duty to insure that the terms of a neutral agreement will not have discriminatory effects is different from the duty to see that no discrimination takes place outside the agreement. But that distinction does not seem meaningful where the alleged union discrimination arises from so important a feature of a collective bargaining agreement as the day-to-day operation of an exclusive hiring hall.
In a case involving discrimination that is given effect by the hiring hall, it might quite as forcefully be said that the union and employer have an “affirmative duty” to prevent discrimination and that the employer cannot use a union “for a shield.”
4. Doctrine of Respondeat Superior
Respondeat superior, a doctrine centuries old, is predicated on the assumption that a master, employer, or principal will be held responsible for the acts of a servant, employee, or agent respectively. The rationale for this view is succinctly expressed by the maxim, qui facit per alium facit per se (he who acts through another acts himself). See, e. g., 3 Am.Jur.2d Agency, § 261, at 626-27 (1962). In the instant case the issue is whether the acts of the union representatives at the hiring hall were, as a matter of law, also the acts of the contractor associations and the contractors who had agreed to this exclusive hiring system. At the outset we are confronted with a line of cases in which some lower courts appear to have held broadly that the concept of respondeat superior does not apply in civil rights cases. Thus we must examine, first, whether a defendant in a civil rights action should be exempt from vicarious liability and then we must consider whether the doctrine of respondeat superior is on its own terms applicable in the present setting.
(a) Analogies to Civil Rights Suits Against Municipalities or Supervisory Personnel
As observed above, it has sometimes been stated axiomatically that respondeat superi- or does not apply in civil rights cases. E. g., Cochran v. Rowe, 438 F.Supp. 566 (D.Ill.1977); Taylor v. City of Selma, 327 F.Supp. 1191 (S.D.Ala.1971); Sanberg v. Daley, 306 F.Supp. 277 (N.D.Ill.1969) (differentiating between monetary and injunctive relief). The courts in these cases, however, and in many of the cases articulating the seeming “axiom,” were confronted with the question of the liability of a municipal entity under § 1983 for acts of its agents. Each case was decided prior to Monell v. Department of Social Services, 436 U.S. 658, 98 S.Ct. 2018, 56 L.Ed.2d 611 (1978), and was therefore predicated on the holding of Monroe v. Pape, 365 U.S. 167, 81 S.Ct. 473, 5 L.Ed.2d 492 (1961), that because municipalities were not persons for purposes of § 1983 and its jurisdictional counterpart, 28 U.S.C. § 1343(3), they were absolutely immune from liability under that section. The rejection of vicarious liability contained in these lower court decisions does not signify a general inapplicability of vicarious liability in civil rights cases which were never subject to the jurisdictional limitations of the Civil Rights Act of 1871, § 1983’s precursor.
Since Monell, of course, municipalities, while still not subject to vicarious liability, may be sued under § 1983 for infringements of civil rights arising from a municipal policy, custom or act. Although overruling Monroe, Monell confirms that in a § 1983 suit against a municipality, respondeat superior is unavailable specifically because of statutory language and legislative history having no bearing on § 1981. In Monell the Supreme Court expressly hinged its rejection of respondeat superior in § 1983 suits against a city on “Congress’ rejection of the only form of vicarious liability presented to it” (i. e., the proposed Sherman amendment) and on “the absence of any language in § 1983 which can easily be construed to create respondeat superior liability.” Id. 98 *410S.Ct. at 2037, n.57. Of course, the rejection of the Sherman amendment imparts no significance for § 1981 whose precursor was the Act of 1866. Mahone v. Waddle, 564 F.2d 1019, 1030-36 (3d Cir. 1977). Furthermore, whereas § 1983 is phrased as a prohibition which might be construed to be applicable only to individual actors,59 § 1981 is phrased so as to grant broadly a positive right and creates no comparable inference. Section 1981 provides:
All persons within the jurisdiction of the United States shall have the same right in every State and Territory to make and enforce contracts, to sue, be parties, give evidence, and to the full and equal benefit of all laws and proceedings for the security of persons and property as is enjoyed by white citizens, and shall be subject to like punishment, pains, penalties, taxes, licenses, and exactions of every kind, and to no other.
In addition to cases involving suits against municipalities, there is also a line of cases under § 1983 (raising the vicarious liability issue) in which supervisory police, prison or other governmental officials are defendants. Typically, respondeat superior is rejected. E. g., Hampton v. Holmesburg Prison Officials, 546 F.2d 1077, 1082 (3d Cir. 1976); Milton v. Nelson, 527 F.2d 1158 (9th Cir. 1976) (§§ 1981-1983); Johnson v. Glick, 481 F.2d 1028, 1034 (2d Cir.) (Friendly, J.) (limiting holding to damage claims), cert, denied, 414 U.S. 1033, 94 S.Ct. 462, 38 L.Ed.2d 32 (1973); Jennings v. Davis, 476 F.2d 1271 (8th Cir. 1973); Adams v. Pate, 445 F.2d 105 (7th Cir. 1971). But see Holland v. Connors, 491 F.2d 539 (5th Cir. 1974).60 The Supreme Court in Rizzo v. Goode, 423 U.S. 362, 96 S.Ct. 598, 46 L.Ed.2d 561 (1976), confirmed that at least with respect to equitable decrees against police department officials, where considerations of comity arose, liability may not be founded on a vicarious liability theory. But as was pointed out by the Third Circuit in Rizzo v. Goode, 506 F.2d 542 (3d Cir. 1974), rev’d on other grounds, 423 U.S. 362, 96 S.Ct. 598, 46 L.Ed.2d 561 (1976), the theory of vicarious liability is inapplicable on its own terms in such situations and is not rendered inapplicable by any peculiarly civil rights related rationale. The Third Circuit Court of Appeals observed:
Respondeat superior is not applicable in its traditional sense since the defendant officials are not employers. Both subordinates and officials are employees of the governmental unit. [Id. at 550]
See, e. g., Kite v. Kelley, 546 F.2d 334, 337 (10th Cir. 1976).
Because of the complete inapplicability here of considerations of federal-state comity, absolute municipal immunity, and immunity of governmental supervisory personnel, the above-listed cases purporting (or seemingly purporting) to exclude respondeat superior liability from civil rights cases properly have no bearing on the present issue.
This conclusion is further compelled by the Third Circuit’s recent decision in Ma-*411hone v. Waddle, supra, reversing the dismissal of a respondeat superior § 1981 claim against the city of Pittsburgh. That decision, pointing out the inapplicability of the legislative history of the Act of 1871 (the precursor of § 1983) makes clear that traditional vicarious liability theories do indeed apply in actions under § 1981. The factual distinction between the present employment discrimination case and the police brutality allegations which were the basis for suit in Mahone cannot alter this basic conclusion. The vitality of Mahone, furthermore, is not diminished by the Supreme Court’s recent decision in Monell.
Based upon the foregoing analysis, I conclude that the plaintiffs’ claim under the “contracts” component of § 1981 may properly be based on a respondeat superior theory, at least to hold the defendant class of employers and the associations injunctively liable for the discrimination by the union in the hiring hall.
(b) The Application of the Doctrine of Respondeat Superior
Under every test of agency the union hiring hall was the agent for two principals — the union and the contractors, with their respective associations. After the contractors delegated authority to the persons operating the hiring hall, can the contractors and associations now be granted absolute immunity from the illegal acts of the hiring hall?
The union hiring hall was, after all, designed to supply the employers with workers and replace the traditional system of direct applications. There is no suggestion here that such delegation is of itself improper, but once it is made the employers relying on it cannot absolve themselves of all responsibility for that delegation even if they are unaware that their collective workforce is the product of intentional discrimination. Nor is it enough that the employers and associations might have liked to avoid a hiring hall system but were forced to acquiesce due to economic considerations. Economic considerations affecting bargaining process raised here by defendants do not provide an excuse from duties incident to employee selection in this case. See Lorillard, supra. Indeed, the employers, particularly through their associations, had the potential collective capacity to be generally conscious of possible discrimination in the hiring hall and to seek to correct it by enforcing the provisions governing the operation of the hiring hall or (at least after May, 1971) by enforcing the non-discrimination provision of the contract. The employers could have taken other actions designed to dissuade the union from a pattern and practice of discrimination in carrying out its duties to supply the employers with operating engineers. The evidence reveals they did nothing whatsoever. The delegation of the employee application process to 542 simply did not effect a withdrawal from reality. See Carey, supra; Lorillard, supra.
There is clear support in Restatement (Second) of Agency for the view which in effect underlies the conclusion reached in Morrison-Knudsen and to some extent in United States Steel, that a relation can exist between an employer or group of employers and an exclusive union hiring hall rendering applicable the rationale of vicarious liability.61 Analogy to Sec*412tion 214 of the Restatement (Second) of Agency is instructive, keeping in mind that an employer can no more easily shed his legal duty under § 1981 than under Title VII not to practice intentional discrimination or produce a discriminatory impact through facially neutral practices in the employment application process:
A master or other principal who is under a duty to provide protection for or to have care used to protect others or their property and who confides the performance of such duty to a servant or other person is subject to liability to such others for harm caused to them by the failure of such agent to perform the duty.62
The duty to see that discrimination does not take place in the selection of one’s workforce must, as was clearly suggested in Carey and Myers, supra, remain with the employer if it is to have full meaning. Employers cannot be permitted, collectively or individually, to use the union as a shield and become oblivious to discriminatory practices whose true harm is revealed in the composition of their own employees. It is still the employer who contracts to hire individuals, even in a hiring hall setting, and that employer should not be free from all responsibility to take steps to alleviate discrimination when found simply because he may be blind to the racial composition of the pool from which workers are referred. Injunctive remedial relief should quite properly be available against such an employer or employers to eliminate discrimination when found.
The contention that each employer’s workforce must be examined individually before even injunctive liability under § 1981 can be found misperceives the thrust of a respondeat superior theory of liability. It is the mass of employers collectively who by originally agreeing to or subsequently adopting the exclusive hiring hall provision, effectuated a collective delegation of the employment application process. Because of this delegation the intentional discrimination of the union was able to work its way broadly into the common workforce of operating engineers. Inasmuch as the delegation of the application procedure was meant to be industrywide within 542’s jurisdiction, and is a product of all those employers agreeing to a single arrangement, those employers must together share responsibility, at least injunctively, for the discrimination by the union. Summarized simply, two elements are necessary to this conclusion and have here been satisfied: (1) the employers delegated an important aspect of their hiring procedure to the union; (2) the union, in effectuating the delegation, intentionally discriminated or, alternatively, produced a discriminatory impact. The acts of the union therefore justify imposition of responsibility upon those employers participating in the original delegation.
Although plaintiffs have failed to show intent to discriminate by the employers as a class, and indeed have been unable to show *413knowledge or notice of discrimination by that class as a whole, plaintiffs have shown that the requisite relationship exists among employers, associations, and union to render applicable the theory of respondeat superi- or, thus making employers and associations liable injunctively for the discriminatory acts of the union.
This decision is not designed to carry with it, however, a presumption of backpay liability specifically against the employers themselves, for questions concerning possible allocation between the intentionally discriminatory union on the one hand and the employers and associations vicariously liable on the other must, by analogy to Title VII, and in the natural order of things, await consideration of the trial court at Stage II. Although the evidence necessary to make this evaluation quite properly has not been presented at Stage I, it has been conceded by plaintiffs at oral argument that among the class of defendant employers there are some whose operating engineer workforces are racially balanced and who might have had no reason to suspect that hiring hall discrimination was occurring or who might have taken steps to avoid discrimination. While these circumstances would not absolve such employers of appropriate injunctive liability for the effects of the hiring hall system, they are factors among other possible factors which may be considered at the damages stage along with the question of allocation. Plaintiffs themselves concede that no liability in backpay should be assessed at this stage inasmuch as that determination may depend on individual issues relating to the culpability vel non of particular members of defendants’ class. Whether such determination can or cannot take place without decertification of the defendant class is an issue not now ripe for decision.
b. Employers and Associations Potential Liability Under § 1985(8)
In the decision of the Third Circuit in Bethel v. Jendoco Construction Co., 570 F.2d 1168, 1172-73 (3d Cir. 1978), the court in the context of an employment discrimination case outlined the elements of plaintiffs’ proof under § 1985(3) as required, by Griffin v. Breckenridge, 403 U.S. 88, 91 S.Ct. 1790, 29 L.Ed.2d 338 (1971):
To come within the purview of 42 U.S.C. § 1985(3), [footnote omitted], Bethel’s complaint was required to allege four matters:
(1) a conspiracy;
(2) for purposes of depriving another . of equal protection or equal privileges and immunities;
(3) any act in furtherance of the conspiracy committed or caused to be committed by a conspirator;
(4) whereby another . . . was injured in his or her person or property or deprived of a right or privilege as a United States citizen.
See also Novotny v. Great American Federal Savings and Loan Association, 584 F.2d 1235 (3d Cir. 1978). Although it is plain that the violations alleged here meet the invidious class-based standard articulated in Breckenridge, it nevertheless seems obvious that no per se or vicarious liability theory could be used to hold a class of employers liable for conspiracy to commit the discrimination practiced by the union. In fact, since the collective bargaining agreement was not itself discriminatory, there was no sufficient proof that as a class the employers agreed to violate equal protection rights or equal privileges and immunities. The conclusion would be the same even if the broad rationale of Dickerson v. United States Steel Corp., 439 F.Supp. 55 (E.D.Pa. 1977), were to be applied. The court there stated that “a party who acquiesces in illegal behavior which could be regulated by contract, either by not pressing for change in the contract or not complaining about an apparent disregard of its provision, has impliedly entered into a conspiracy.” Id. at 67. Not even acquiescence of the whole class of employers in the sense of a conscious toleration of the discrimination of the union has been shown. Absent such class-wide proof, no § 1985(3) claim could be upheld even on this expansive theory. Furthermore, to the extent that individualized *414proof against particular employers may be necessary to establish a § 1985(3) claim, it is not appropriately raised against a class of defendant employers.
The liability of the associations under § 1985(3) raises a closer question because of their involvement in dramatically misstating the minority population of 542 in avoiding application of the Philadelphia Plan. Nevertheless, despite this reckless disregard for the accuracy of these figures and other proof of inaction, I do not on the present record find actual knowledge of the formal error in figures or of union discrimination generally.
C. Appropriateness of the Defendant Class as to Section 1981 Claims
Having determined that in this factual setting the scope of § 1981 permits per se or vicarious liability at least as to equitable relief on the part of all employers who delegated exclusively to 542 the role of employee-applicant selection, much of the basis for objection of defendant nonparties to the defendant class is dissipated, for there are no individualized issues which might be overlooked. It is nevertheless appropriate to consider in detail the applicability of Rule 23(a) and (b), the issue of standing of the plaintiff class to assert claims broadly against the defendant class, and the issue of personal jurisdiction over the defendant class members.63
1. Rule 23(a)
First, the requirements of rule 23(a), considered above in discussing the composition of plaintiffs’ class, must be analyzed with respect to the defendants’ class. As has been pointed out, rule 23(a) contains four basic requirements: numerosity; commonality of questions of law or fact; typicality of claims and defenses; and the adequacy of the class representative to protect the class interests. While numerosity is not in question, the remaining requirements have been assailed; they are best discussed upon examination of the representative’s adequacy and a recapitulation of the claims against the defendant class.64
Defendant Glasgow, Inc., is one of the largest employers of operating engineers in 542’s jurisdiction. It is a member of the CAEP and is a party to the collective bargaining agreement providing for an exclusive hiring hall. It regularly receives referrals of operating engineers from 542’s hiring hall. Also named as defendants are the trade associations which have as members employers of operating engineers in 542’s geographical range. These associations are defendants not only because of the composition of their membership, but because of their administrative capacity as associations to serve as negotiators or spokesmen for their members. Both named associations and named defendant employer are represented by the same counsel, whose abilities to conduct this litigation cannot seriously be questioned.65
*415Focusing on defendant Glasgow’s representative status, Wetzel, supra, requires us to consider whether Glasgow and other members of defendants’ class are in antagonistic positions with respect to the subject of the claims against them. It must be concluded that they are not, although non-parties have asserted numerous arguments against such a conclusion. Nonparties have urged that it is in Glasgow’s personal interest to keep as many employers in the suit as possible so as to dilute potential liability. As an indicium of this antagonism nonparties allege Glasgow’s failure to oppose certification of the defendant class and to respond to a memorandum, requested of plaintiffs by the court, on the appropriateness of the defendant class certification. Nonparties have also argued that historical antagonism between national and local contractor-employers makes joining them under the representation of a local contractor and associations improper.
Named defendants did oppose the defendant class certification as proposed by plaintiffs at oral argument on February 18,1972, though they did not reject the total concept of it. But more importantly if this court now otherwise accepts the proposed defendant class as proper under rule 23, having the additional benefit of all the objections of nonparties, the level of named defendants’ opposition to defendant class certification seems virtually meaningless.
Concerning the theory that a named defendant would prefer to be sued as a member of a large group rather than alone and that such a preference defeats its adequacy of representation, not much need be said. Taken to its logical conclusion the theory could prevent every defendant class action. The real question is whether representation by a named defendant of a class of defendants would result in a less than full and aggressive representation. Defendant Glasgow had absolutely nothing to gain by defending the action poorly since that could create liability in itself, even if liability were broadly distributed. And so long as the terms by which liability would be imposed upon a named defendant employer are exactly the same as the terms upon which any other defendant employers might be held liable, there is sufficient assurance that the named defendant will represent absent defendant class members adequately. The elements of § 1981 which plaintiffs had to prove against the defendant class are elements which Glasgow and the associations had a definite interest in rebutting to avoid individual liability — union discrimination and the entrustment of employment application processes to the hiring hall by agreement with the union. As to these issues there can be no distinction between national and local contractors. Even if some employers among the class may have used C or D branch members whose referral was not regulated by the contract with the construction branches, that does not make them different with respect to the issues relating to construction branch employment. There is thus no legal antagonism with respect to the only basis for defendant class liability.
Hand in hand with the lack of legal antagonism, and hence the adequacy of the named defendants, commonality and typicality requirements are fulfilled. As has been stated, the issues of fact and of law are precisely the same for all employers at this liability stage. With respect to the theory of per se liability adopted herein there is likewise no individualized defense. It is, to repeat, not now properly considered whether the representation or certification at the liability stage may continue into Stage II.
2. Rule 23(b)
Nonparties additionally attack defendant class certification on the ground that certification under 23(b)(2) or (b)(1) is improper in general terms and that (b)(3), which permits a class member to opt out after notice is the only permissible method of certification. Notwithstanding these arguments, the defendant class does indeed fall within rule (b)(2).
Rule 23(b)(2) provides that a class may be certified where
*416the party opposing the class has acted or refused to act on grounds generally applicable to the class, thereby making appropriate final injunctive relief or corresponding declaratory relief with respect to the class as a whole .
It would be somewhat incongruous if a class of plaintiffs must have “acted or refused to act” on grounds generally applicable to a class of defendants in order to validate the defendant class, for that would place the power of creating a defendant class into the hands of a possibly obstinate and unreasonable plaintiff. 3B J. Moore, Federal Practice ¶ 23.40 at p. 103 (Supp.1977-78). But where, according to the allegations and proof, a class of plaintiffs is forced, if they wish to be employed as operating engineers, to act with respect to the defendant class of employers in one and only one way, the requirements of 23(b)(2) are met.
It must be remembered that rule 23(b)(2) or any part of rule 23 is not by its terms applicable only to a plaintiff class certification. Indeed, the language in 23(b)(2), “the party opposing the class,” on its face clearly indicates that the section was deliberately cast to include defendant class action suits. Numerous cases have upheld 23(b)(2) defendant class certifications in a variety of contexts. E. g., Redhail v. Zablocki, 418 F.Supp. 1061, 1066 (E.D.Wisc.1976) (three judge court), aff’d on other grounds, 434 U.S. 374, 98 S.Ct. 673, 54 L.Ed.2d 618 (1978); Thompson v. Board of Education of Romeo Community Schools, 71 F.R.D. 398, 402 (W.D.Mich.1976); Bradford Trust Co. v. Wright, 70 F.R.D. 323 (E.D.N.Y.1976); Paxman v. Wilkerson, 390 F.Supp. 442, 447-48 (E.D.Va.1975); Gibbs v. Titelman, 369 F.Supp. 38, 52-53 (E.D.Pa.1973), rev’d on other grounds, 502 F.2d 1107 (3d Cir.), cert, denied, 419 U.S. 1039, 95 S.Ct. 526, 42 L.Ed.2d 316 (1974). Washington v. Lee, 263 F.Supp. 327, 330 (M.D.Ala.1966), aff’d, 390 U.S. 333, 88 S.Ct. 994, 19 L.Ed.2d 1212 (1968); see Note, Defendant Class Actions, 91 Harv.L.Rev. 630, 633 (1978). There are, however, cases which hold or suggest otherwise. Mudd v. Busse, 68 F.R.D. 522, 529 (N.D.Ind.1975), appeal dismissed, 437 F.Supp. 505 (1977); cf. United States v. T. I. M. E.-DC, Inc., 335 F.Supp. 246 (N.D.Tex. 1971) (holding, alternatively, that where various union locals were sought to be joined as defendant class, court lacked in personam jurisdiction over them so long as alteration of contract was envisioned).
While among the cases permitting a (b)(2) defendant class certification Redhail and Gibbs were suits seeking to enjoin a class of state officials from enforcing a generally applicable state statute alleged to be unconstitutional, Thompson and Paxman were wider ranging suits seeking to enjoin, respectively, school boards and school board members for policies alleged to discriminate on the basis of sex. Thompson upheld a 23(b)(2) defendant class consisting of “all school boards in the state of Michigan which . . . have treated or now treat pregnancy related disability differently than other temporary disabilities . . .” Id. at 402. Similarly, Paxman upheld a class of school board members whose districts applied discriminatory maternity leave policies, even though maternity policies could vary in their precise terms from district to district. Paxman and Thompson suggest a very practicable solution: if that which makes a practice illegal is common to the defendants, there can be no cognizable objection to the concept of a (b)(2) defendant class certification. Where, of course, issues are not identical or closely parallel, there would be a serious argument against binding an absent defendant under res judicata principles to the result of a litigation in which he did not participate and of which he was not entitled to notice. Such a circumstance, not presented here, might cause some reluctance to apply what appear to be the plain terms of 23(b)(2), perhaps on the theory that congressional intent could not be found to support the result. But where the litigable issues are identical or nearly so, the efficacy of proceeding by a defendant class action logically outweighs the arguments favoring multiple individual suits and fits tightly with the overall scheme of 23(b)(2). Where such identity occurs there is no greater reason to deny classwide res judicata application against defendants than against plaintiffs.
*417The importance of (b)(2) certification in a case such as this is heightened because of the impossibility of obtaining make-whole injunctive relief from employment discrimination against any particular defendant. The hiring hall system is a method of distributing the labor resource throughout relatively large areas in which various contractors may be intermittently engaged. As a practical matter, given the allegations in this case, if plaintiffs’ class (or an individual plaintiff for that matter) were limited to suing individual defendant employers there could be no assurance that a mandatory remedial level of hours and wages could ever result, or that strict adherence to the rules of the hiring hall would be broadly maintained by employers, or that employers as a whole could be obliged to take steps to assure an awareness of the levels of minority employment in 542’s jurisdiction. Each of these matters concerns relief which may be necessary to each discriminatee, for operating engineers in a hiring hall setting cannot regulate or foresee those employers to whom they may be referred.
3. Standing
The question remains, however, whether named plaintiffs have standing to sue a class of defendants whose members include employers by whom no named defendant was employed as well as employers operating outside District I, the place of residence of all named plaintiffs. This issue is appropriately raised in a discussion of the scope of defendants’ class because a lack of standing by named plaintiffs would affect defendant class certification.
Nonparties rely heavily upon two decisions, La Mar v. H & B Novelty & Loan Co., 489 F.2d 461 (9th Cir. 1973) and Weiner v. Bank of King of Prussia, 358 F.Supp. 684, 694-701 (E.D.Pa.1973), which, they assert, instruct that a decertification must here be ordered. Neither case, however, is decisive of the issues now raised.
La Mar was a consolidation of two actions. One was a suit by a plaintiff class represented by a single named plaintiff, La Mar, against a defendant class of pawnbrokers. The other was a suit, again by a single named plaintiff, Kingsling, on behalf of a class of plaintiffs against eight airline companies. Both cases involved alleged overcharges of nominal amounts stemming from practices common to the respective groups of defendants. La Mar alleged personal injury from only one pawnbroker and Kingsling alleged injury from only two air carriers. Although declining to decide that the named plaintiff in each action lacked standing to sue those defendants who caused him no injury, the court made use of a standing rationale in determining that the named plaintiffs lacked the requisite typicality, stating that “the representative plaintiff’s cause of action is against a defendant unrelated to the defendants against whom the cause of action of the members of the class lies.” Id. at 465.
Even accepting the rationale of La Mar on the standing issue, it still appears that La Mar is inapplicable in the present setting. The court specifically distinguished, for instance, the case of Washington v. Lee, 263 F.Supp. 327 (M.D.Ala.1966), in which two black prisoners, alleging segregation, were permitted to sue a defendant class of all sheriffs and wardens of city and town jails in Alabama, even though the named plaintiffs were not themselves alleging injury by all such sheriffs and wardens. The La Mar court was persuaded that
Aside from the somewhat broad and accommodating concept of standing in civil rights cases, [footnote omitted] it is also true that the defendants were officials of a single state and its subordinate units of government. Their legal relationship distinguishes them from the defendants the plaintiffs La Mar and Kingsling seek to envelop in their class action. [Id. at 469-70]
The court later registered its belief that mere common commercial practice does not constitute a sufficient “legal relationship.” It is thus fair to conclude that the La Mar result was not intended to reach civil rights cases and especially not civil rights cases in which the members of a proposed defendant class were engaged in a joint, not merely *418comparative, undertaking such as sharing the hiring hall services of a single local union under the same terms and conditions. This distinction may be particularly significant where more than nominal relief is sought.
Weiner v. Bank of King of Prussia, a case decided squarely on Article III standing grounds, held that a single plaintiff having dealings with a single bank cannot assert a National Bank Act or Truth in Lending Act claim against all banks in the court’s jurisdiction alleging failure to properly disclose interest rates. It is important to understand that in this commercial suit the group of defendant banks had no legal relationship whatsoever (contractual or otherwise) to one another. Weiner therefore does not resolve the questions presented here.
The Third Circuit’s decision in Kauffman v. Dreyfus Fund, Inc., 434 F.2d 727 (3d Cir. 1970) (Aldisert, J.), served as the stated basis for the Weiner decision and was also quoted in support of the Ninth Circuit’s La Mar decision. Kauffman, like Weiner, was a standing case in which a plaintiff holding shares in four mutual funds was held, inter alia, to lack standing to sue 61 other mutual funds for like practices in a derivative-type action alleging securities violations and illegal antitrust conspiracy to fix fees and to monopolize. Inasmuch as the derivative action as defined in rule 23.1 requires that one who is not a shareholder in a corporation cannot bring a derivative action in its behalf, Kauffman’s holding is very narrow indeed. The court tellingly observed that:
A secondary right is limited; it is co-extensive with plaintiff’s need as a shareholder to protect an interest which allegedly has been impaired. And the very limitation that inheres conceptually in a derivative action is the bar which defeats appellee’s attempt to qualify as the class representative for the other funds. [434 F.2d 737]
The court did not rule out the possibility that, had the plaintiff’s right against the four funds of which he was a shareholder been primary, he could properly have sued the other 61 mutual funds.
To be sure, if the fund itself sought to bring a class action in behalf of other funds similarly situated, there would be no theoretical barrier prohibiting such a class action. [Id.]
Weiner’s stated explanation for this distinction in Kauffman is the existence in Kauffman of an antitrust conspiracy count, but the Court of Appeals did not so limit its language. Kauffman certainly does not rule out the availability of the defendant class device where named plaintiffs have a true direct claim or cause of action against at least one of those defendants and all issues are common. See La Mar, supra at 468-69.
Even if La Mar and Weiner were viewed as correct statements of the standing doctrine in class actions, and this need not here be addressed, the present action would not be materially affected. There is in this case a single entity, Local 542, which was alleged and proved to have undertaken a pattern and practice of intentional discrimination (and to have acted in such a way as to have caused a substantial discriminatory impact). The named plaintiffs clearly have standing to sue Local 542 even though they are all residents of District I. The five district divisions in 542 are, moreover, directly related to each other and administratively under the direction of one Business Manager. In some cases job assignments may be obtained interdistrict, and the same collective bargaining agreement applies in all districts, though each maintains its own out-of-work hiring hall listings.
The two asserted standing defects, to repeat, arise from (1) the failure of named plaintiffs to show that they made direct applications to or had direct dealings with all the defendant employers and (2) the representative plaintiffs’ common residence in District I. It must be reiterated that the governing entity of 542 consists of a single executive board comprised of 542 officers and elected representatives from each district. Further, because of the nature of 542’s hiring hall system any given minority listee is in a pool of applicants to all and any employers who may request a referral in one of the five districts.
*419Employers using this hiring hall system rely on a constant presence of some available workers in five convenient geographical areas, being aware also that the exact composition of those workers is changing constantly as referrals are made and as individual jobs are completed. It is precisely because of this method, supplanting the direct application process, that employers may be said to have had dealings with the members of the plaintiffs’ class and its representatives even though no plaintiff may have applied to any employer directly and even if only a small number had actually been employed.
In one sense, of course, plaintiffs might have been able to limit their proof to discrimination in District I or to discrimination personally against themselves, avoiding a class action suit altogether. But Title VII permits suit against the labor organization alleged to be the source of employment discrimination, not just a piece of it. Indeed, Title VII by its very nature, envisions class action litigation. In this light, while named plaintiffs might have acted otherwise, they had standing to sue 542 as a whole on a class action basis. The only cognizable claim which plaintiffs have against the class of defendants results from the denial of equal employment rights through the hiring hall system which, to the extent articulated above, was the responsibility of all class defendants. Named plaintiffs do therefore have a personal stake in the case against all defendants. See Baker v. Carr, 369 U.S. 186, 204, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962). It follows that those employers dealing with the exclusive hiring hall system are rightly joined in this action as parties accountable for 542's Title VII or 1981 violation, even if the suit against the employers is based on 1981 alone. La Mar, Weiner, and Kauffman certainly are not authority for a contrary conclusion on these facts.
Also, as the La Mar court observed, in civil rights cases plaintiffs’ standing is to be interpreted as expansively as Article III permits. Hackett v. McGuire Bros., Inc., 445 F.2d 442 (3d Cir. 1971), states:
The national public policy reflected both in Title VII of the Civil Rights Act of 1964 and in § 1981 may not be frustrated by the development of overly technical judicial doctrines of standing or election of remedies. If the plaintiff is sufficiently aggrieved so that he claims enough injury in fact to present a genuine case or controversy in the Article III sense, then he should have standing to sue in his own right and as a class representative. [Id. at 446-47]
There are sufficient, if unique, connections among defendant employers, the union, and the named plaintiffs so that named plaintiffs’ standing to sue the defendant class on behalf of all plaintiffs on a vicarious liability theory must be upheld. Named plaintiffs have a claim against all defendants and hence have a stake in the outcome of the litigation sufficient to assure concrete adverseness and at the same time afford to rule 23 a scope commensurate with its purpose.
4. Personal Jurisdiction Over the Defendant Class
The question of personal jurisdiction over members of the defendant class is quite separate in theory from the Article III standing objections discussed herein. The argument raised by nonparties is essentially that due process considerations protect unnamed defendants from liability, particularly monetary liability. However, as the Supreme Court stated in Hansberry v. Lee, 311 U.S. 32, 61 S.Ct. 115, 85 L.Ed. 22 (1940):
members of a class not present as parties to the litigation may be bound by the judgment where they are in fact adequately represented by parties who are present. . . . [Id. at 42-43, 61 S.Ct. at 118-119]
In this case the class of defendants was adequately represented. Particularly because the only basis for liability of the defendant class was vicarious liability under § 1981 and because only injunctive liability has been decided at this stage, there is no room for argument that the named defend*420ant Glasgow, Inc., was unrepresentative. I must also reiterate that notice was provided to the unnamed members of the defendant class by certified mail two years prior to the trial in this case. In these circumstances and given full compliance with rule 23 there seems little doubt that through the defendant class device personal jurisdiction over the defendant class was achieved.
United States Circuit Judge sitting by designation.
. Those trade associations are: Contractors Association of Eastern Pennsylvania; General Building Contractors Association, Inc.; Pennsylvania Excavating Contractors Association; and United Contractors Association.
. Although notice was provided out of an abundance of caution to protect the interests of fairness, no unnamed members sought to intervene in this action.
. This agreement was negotiated by the CAEP and GBCA on behalf of its members. Not all employers were members; however, the defendant class is limited to all employers who have adhered to the agreement.
. The basic terms appear consistently in the bargaining agreements having a bearing on this case.
. Other restrictions pertaining to referral are the three refusal rule and the 90 day no-recall rule. An operating engineer who refuses without excuse an offer of employment three consecutive times is to be placed at the bottom of his out-of-work list. No employer is to recall outside the referral system after 90 days of separation.
. Prior to May 1, 1971, Art. II, § 2(d), expressly mentioned only discrimination on grounds of non-membership in the union. The provision read as follows:
(d) The selection of applicants for referral to jobs shall be on a non-discriminatory basis and shall not be based on, or in any way, affected by, Union membership, by-laws, regulations, constitutional provisions or any other aspect or obligation of Union membership, policies or requirements, except as outlined in this Agreement.
After May 1, 1971, a provision was added:
No employee, or applicant for employment, shall be discriminated against by reason of race, religion, color, or national origin. [Art. II. § I-]
. “Men” is used here because the record establishes that 542 has never had women members of the union.
. U-261.
. Transfers from other locals can also be achieved once a clearance card is obtained.
. 30 Fed.Reg. 12319.
. The Secretary of Highways of the State of Pennsylvania objected to the demand for a “manning table” asserting, inter alia, that such a demand contravened civil rights law. This issue is not now before us.
. In a letter from Martin Vigderman, Esq., to Bennett Stalvey, dated January 8, 1969, the number of minority members was stated to be about 350 of a total of about 5500 members.
. Homer Dawson, in a letter to John L. Wilks of the OFCC dated December 6, 1971, stated that in a 1968 report to the EEOC the union expressed the result of a questionnaire on minority membership. The union reported 193 minority members of 2845 responses. A total of 6000 questionnaires were stated to have been mailed out.
. The union never called Holland to testify to refute those claims even though he was working as an operating engineer during most of the time on a construction project in this very courthouse and even though, on occasion, he was a spectator at the trial.
. The union, associations and contractors were also obligated as a result of their “Affirmative Action Program” to upgrade their training opportunities with an emphasis on minority participation. Although the “Affirmative Action Program” succeeded in diverting application of the Philadelphia Plan, it did not, even at this basic level, provide the assured attention to minority upgrading. Operation Stepping Stone, an upgrading project pursuant to the “Affirmative Action Program,” gave minimal attention to the purpose of minority upgrading. No special efforts were made to notify minority operators or to describe the program as one which was to emphasize upgrading among minorities. As to four minority applicants to the program who did not participate, the record reveals no explanation. A total of six minorities out of 96 participants in the two-session operation spent at least some upgrading time. Only three minority individuals out of 89 completed the program.
. Although certain exhibits containing letters, findings and decisions pertinent to the administrative disposition of this case before the EEOC were excluded for substantive purposes, they must be considered purely in addressing the issue of the existence of and the extent of jurisdiction. These legal issues will be discussed below.
. The record does not make reference to the sending of a right to sue letter to Muchison.
. The adjustment is made by subtracting the withdrawals and reinstatements from the group considered to have entered in 1975.
. Union’s Exhibit 254 indicates that there were only two D branch members between 1972 and 1975.
. This last figure may actually be 18.9% based on calculations from Plaintiffs’ Exhibit 253(f)-(h) indicating the rate of minority entry from 1972-74.
. The possibility of an interdistrict effect on these differentials was also explored by analyzing District I according to the same standards. District I holds 85% of 542 but only about 50% of white members. Siskin concluded that the District I analysis produced the “same general result.” P-168, Rpt. 7 at 12. This analysis produces results reasonably comparable to those tor the union generally: whites of any branch and list group work more hours than minorities of comparable status. Notably, the rate of pay differential between minority and whites in District I is greater for A branch, B branch (Group 1), and RA, but the 1972 study roughly confirms Siskin’s earlier estimate of the differential.
. It is also noteworthy that the process of making referrals based on skill was not an absolutely formal one and did not, according to Mr. Ciavaglia, the District I dispatcher, arise “that often.” Tr. 2035. Skill reversals were made by Ciavaglia according to his memory and work record notations. The work records are a wholly unreliable source of skill information.
. This is described more fully under the heading JATC below.
. Dent became a parent body member in August of 1973.
. It must be noted that despite Johnson’s skills on several types of machinery and although he informed the union of these skills, no machinery skills were listed on the work record.
. Benjamin’s phone number had not changed since he became a registrant. He also testified that someone was generally at home even when he was not.
Benjamin also testified that he informed Ciavaglia at the time he was issued a green book that he could operate several specific types of machinery; however, no skills were listed on his work record.
. Working the 8:00 a. m. — 4:30 p. m. shift, Allen reported that he observed Local 542 operating engineers on the job, numbering 85 to 100 men. Of that number, no more than 5, at any given time, were black — and they were working primarily as oilers, not operators.
. There is, of course, no legal requirement that representation in the union be precisely proportionate to the labor pool representation. The evidence in this case indicates in general a gross statistical disparity which shows intentional discrimination but which is not of its nature irrefutable.
. Wachter states in his report (Union’s Exhibit 428 at 46^47): “It is clear that on average the percentage of minority applicants to the [RA] system far exceeds the minority representation in the community.”
. Wachter, of course, does not specify any particular percentage by which he theorizes the labor pool should be reduced for any of the above factqrs. Furthermore, the abilities needed for purposes of general entry at low priority levels are of the sort that “many persons possess or can fairly readily acquire.” Hazelwood, 97 S.Ct. at 2742 n.13. This is not to say that all persons who enter are unskilled, but that numerous entry routes have existed during times pertinent here (or still exist, though they have recently been reduced) which do not require any level of skill as an operating engineer. The gross disparities in membership create an infer*372ence of purposeful discrimination. That is supplemented in this case with non-statistical proof as indicated above and with other statistical proof which favors the inference created by the original membership disparity. In the course of time, given the existence $f numerous unskilled entry methods, even if there were the slightest evidence that minority individuals were disproportionately unskilled to work as operating engineers, one would expect the percentages to adjust themselves to Siskin’s labor pool calculation, absent discrimination, so that such a gross disparity would not exist.
. While I do not disagree with Dr. Siskin’s alternative 13.9% minority labor pool figure (reducing total labor pool by eliminating those who have a greater than high school education), I have not relied upon that 13.9% figure in the analysis of plaintiffs’ case. Instead, I have used the conservative 11.5% measure. (Most of plaintiffs computations were based on this figure.) The 13.9% figure, however, seems realistic in the sense that because of a generally lower formal educational attainment among minorities, the proportion of the minorities available for and interested in operating engineer work may be greater than that of whites who because of proportionately greater formal educational attainments might have additional opportunities.
. Dr. Wachter has also stated that “In no case should a finding of [statistical] under-representation alone be taken as proof of discrimination.” This underlying view is at odds with the Court’s conclusion in Hazelwood that “[w]here gross statistical disparities can be shown, they alone may in a proper case constitute prima facie proof of a pattern or practice of discrimination.” Id. 97 S.Ct. at 2741. I therefore must also reject this too-broadly stated view.
. Contractor’s Exhibit 6 at 6 (emphasis added).
. Such a comparison does, of course, provide a possible avenue for rebuttal evidence as is explained in Hazelwood.
. Wachter points to the relative equality in the percentage of minority and white entrants into the union from the registrant program as demonstrating that no referral discrimination occurred. Apart from Wachter’s selectivity about when to bring up this relative equality of entry, in the absence of evidence of the reasons for the disproportionately large number of withdrawals by whites from the registrant program, no such inference seems appropriate.
. The 1969 71 study estimated the out-of-work list groupings. The estimates treated whites and minorities in like categories of hours worked.
. At least one of the minority individuals, Lucius Randolph, was originally listed as having passed the BES test but was later told that he failed. This resulted when the JATC had the BES test scores checked for those who scored low on the JATC test. Plaintiffs contend that when the JATC realized the high number of minorities which had passed the BES test it “upped” the passing grade, as it were, guided by the belief that commensurate with past experience proportionately more minorities than whites would be excluded. Approximately twenty persons were deemed unqualified as a result of these changes, but there is no admissible evidence on how many were minorities.
For specific instances of the treatment of minority individuals by the JATC, see the narration of testimony supra.
. A study of the 1969 RA class showed that minority RAs worked more hours in 1972 than did white RAs on the average. There were only 3 minorities among a total of 49 workers in this study and 1972 was the only year for which this was true. Other analyses by Siskin of individual classes (1970-1972) reveal, consistently, fewer hours worked by RAs.
. As some of the evidence that educational requirements were inappropriate there is the fact that in the 1966 class of A branch entrants (prior to the educational qualifications), 85.1 % of those having less than a full high school education successfully completed the RA program as opposed to 86.5% of those with full high school educations.
. Among the 289 minorities who took that test, 20.4% failed it. Among the 1937 whites who took the test, 10% failed.
. Only part of 1975 was studied.
. The data for these years excludes those who were referred to jobs in which less than three operating engineers were engaged. Siskin assumes the sample does not affect randomness and hence assumes that minorities were not assigned disproportionately to such small jobs, or for disproportionate lengths of time. With this qualification his analysis bears a .05 level of significance.
. Because of my decision on Title VII liability and liability under § 1981 and solely because those provisions permit complete and comprehensive relief, although as yet remaining to be determined, I do not here decide: (1) the issue of legal liability of the union under the National Labor Relations Act, 29 U.S.C. § 141 et seq. (1970) for breach of its duty of fair representation; or (2) the claim of a private action for discrimination in programs receiving federal funding under Title VI, 42 U.S.C. § 2000d et seq. (1970).
. Rule 23(a) provides as follows:
(a) Prerequisites to a Class Action. One or more members of a class may sue or be sued as *387representative parties on behalf of all only if (1) the class is so numerous that joinder of all members is impracticable, (2) there are questions of law or fact common to the class, (3) the claims or defenses of the representative parties are typical of the claims or defenses of the class, and (4) the representative parties will fairly and adequately protect the interests of the class.
. It is apparent also that, although the Court focused on the class membership issue in its discussion, citing cases delineating the Article III standing doctrine, it also considered named plaintiffs’ stipulation and the proof of their lack of qualifications in concluding that rule 23(a) was not satisfied. See id. 419 U.S. at 401-02, 95 S.Ct. 553. The Court also added that the plaintiffs’ failure to move for class certification prior to trial and the existence of a conflict between the plaintiff’s complaint and a union vote of class members created indicia of inadequacy.
. I adopt the six-year period for the § 1981 claim for reasons expressed in my opinion in Collier v. Philadelphia Gas Works, 441 F.Supp. 1208 (E.D.Pa.1977). See Davis v. United States Steel Supply, 581 F.2d 335 (3d Cir. 1978) (applying six-year statute of limitations by analogy to 12 P.S. § 31 in a case of discriminatory discharge).
The period by which the class of those having a Title VII claim is defined began as of March 19, 1969, 90 days prior to the time of the filing of an EEOC charge by named plaintiffs Williams and McKay. The recertification incorporates this limitation. I reject use of the Ximines charge for this purpose in the absence of specific authority on point and because as a practical matter the availability of relief under § 1981 encompasses the relief available under Title VII. This issue might become important if § 1981 liability is struck out of the case on an appeal.
. Teamsters involved the issue whether city drivers who were discouraged from transfer to “line driver” status could obtain relief without having actually made an application. The Court stated rather broadly:
The denial of Title VII relief on the ground that the claimant had not formally applied for the job could exclude from the Act’s coverage the victims of the most entrenched forms of discrimination. Victims of gross and pervasive discrimination could be denied relief precisely because the unlawful practices had been so successful as totally to deter job applications from members of minority groups.
431 U.S. at 367, 97 S.Ct. at 1870.
. Sections 2000e -2(c) and (d) provide:
Labor Organization Practice
(c) It shall be an unlawful employment practice for a labor organization—
(1) to exclude or to expel from its membership, or otherwise to discriminate against, any individual because of his race, color, religion, sex, or national origin;
(2) to limit, segregate, or classify its membership or applicants for membership, or to classify or fail or refuse to refer for employment any individual, in any way which would deprive or tend to deprive any individual of employment opportunities, or would limit such employment opportunities or otherwise adversely affect his status as an employee or as an applicant for employment, because of such individual’s race, color, religion, sex, or national origin; or
(3) to cause or attempt to cause an employer to discriminate against an individual in violation of this section.
(d) Training programs
It shall be an unlawful employment practice for any employer, labor organization, or joint labor-management committee controlling apprenticeship or other training or retraining, including on-the-job training programs to discriminate against any individual because of his race, color, religion, sex, or national origin in admission to, or employment in, any program established to provide apprenticeship or other training.
. The court in Jenkins was divided on the application of the rule where a woman alleged that she had been discriminated against because of her Afro hair style and where the charge stated that the employer accused her of being “a leader of the girls on the floor.” The majority concluded that these statements in the charge properly raised both racial and sex discrimination issues even though the box denoting sex discrimination was not checked. The dissent felt that the statements did not suggest a pattern or practice of discrimination and that the core of the plaintiffs claim was that she was discriminated against due to her hair styling.
. Local 542 has cited EEOC v. Allegheny Airlines, 436 F.Supp. 1300 (W.D.Pa.1977). In that case a male job applicant charged to the EEOC that only females were hired as flight attendants. The Commission also investigated other practices, namely, the failure to hire and promote women and blacks to executive and technical positions. When the sex discrimination charge could not be conciliated in the Commission’s view, conciliation was concluded and suit was filed by the EEOC alleging racial and sex discrimination in hiring and promotions. The court held the scope of the district court action was too broad, not because of the terms of the original charge but because the conciliation attempts were in effect limited to the allegations of that charge. (Indeed, the court observed that the Commission had not acted unreasonably in pursuing the investigation of hiring and promotion discrimination.) It is sufficient to note for present purposes that this decision has no bearing on the present action because it only deals with limitations on the power of the EEOC itself to sue after insufficient conciliation attempts.
. This conclusion does not run afoul of § 2000e-2(h), which states:
it shall not be an unlawful employment practice for an employer to apply different standards of compensation, or different terms, conditions, or privileges of employment pursuant to a bona fide seniority . . . system . . . provided that such differences are not the result of an intention to discriminate because of race ... or national origin .
Teamsters held this provision to bar the conclusion that a seniority system which has the effect of perpetuating pre-Title VII discrimination but which is not itself discriminatorily motivated violates Title VII. Adherence to the experience-graded referral system, which on its face resembles a form of seniority division, has been sporadic, leaving much room for arbitrary decisionmaking. Furthermore, § 2000e-2(h) is not infringed by a decision that the union has committed a disparate impact violation of Title VII, because plaintiffs do not now argue that the referral system on its face constitutes a discriminatory practice.
. Because of the JATC’s participation in the overall intentional discrimination of the union, there is no need to discuss its legal liability separately. The JATC is liable as the union is liable.
. In Ferguson the court stated that by making the union its exclusive hiring hall it “made Goodman [the union’s business agent] its agent in determining . . . eligibility for employment and by doing so became liable for any unfair practices committed by Goodman in connection with his hiring or refusing to hire.” Id at 208. Although there is no discussion of the employer’s knowledge of the specific discriminatory referral revocations involved in Ferguson, the court also described the employer’s involvement as “condoning” or “acquiescing” in encouragement of union membership.
. The court reversed an order by the NLRB requiring the employer to reimburse all illegal union dues. The court pointed out that this would impose an unfair burden on the employer and permit the union to retain the funds illegally received.
. The court in United States Steel also concluded that the employer participated in an illegal discriminatory arrangement although it did not expressly find that the employer had knowledge of any discriminatory provision. The NLRB decision, 122 NLRB No. 155, reveals that an employer-manager tried to have nonunion workers hired by the master mechanic. The NLRB nevertheless found the arrangement to be “in effect” an agreement to operate under closed shop conditions. Id. at 1325. In the present case, although there has been no allegation that the bargaining agreement is itself discriminatory and hence although there is some possibility that there was no analogous “illegal arrangement,” the court’s analysis of the agency issue seems immediately instructive.
. The Second Circuit’s opinion summarized the affirmative action order, 542 F.2d at 589:
MCA was enjoined from further discrimination and ordered to maintain up-to-date records of available work, to submit an affirmative action program, to use its best efforts to provide apprentices with 1,750 hours of work per year and to maintain an employment register.
. The basis for this holding was the interpretation given to 42 U.S.C. § 2000e-2(h), which provides in part as follows:
Notwithstanding any other provision of this sub-chapter, it shall not be an unlawful employment practice for an employer to apply different standards of compensation, or ditferent terms, conditions, or privileges of employment pursuant to a bona fide seniority system, . provided that such differences are not the result of an intention to discriminate because of race . or national origin .
. A recent Fifth Circuit decision in the area is James v. Stockham Valves & Fittings, 559 F.2d 310, 353 (5th Cir. 1977) (Wisdom, J.). In this case the court was faced with a seniority term, not bona fide, which impacted discriminatorily. The court also found discriminatory purpose on the part of the employer. But the court remanded on the question of the union’s liability, notwithstanding that the union had agreed to the collective bargaining contract. The court was obviously influenced by the union’s strong efforts to eliminate the discriminatory provisions. Those circumstances are not present here, however, for the employers and associations have taken no meaningful steps to see that discrimination is resisted.
. 42 U.S.C. § 1983 provides:
Every person who, under color of any statute, ordinance, regulation, custom, or usage, of any State or Territory, subjects, or causes to be subjected, any citizen of the United States or other person within the jurisdiction thereof to the deprivation of any rights, privileges, or immunities secured by the Constitution and laws, shall be liable to the party injured in an action at law, suit in equity, or other proper proceeding for redress.
. Another less familiar setting in which vicarious liability claims have been raised is in § 1983 or § 1985 cases in which private parties are alleged to have participated with state officials. Two such cases are Draeger v. Grand Central, Inc., 504 F.2d 142 (10th Cir. 1974), and Croy v. Skinner, 410 F.Supp. 117 (N.D.Ga. 1976).
In Draeger an off-duty police officer was serving as a security guard for a private store. Although off-duty, the policeman still had authority to make police arrests. Considering the question whether the employer-store was liable for the alleged illegal arrest by their security guards, the court held that respondeat superior was not applicable. However, in Croy v. Skinner, 410 F.Supp. 117 (N.D.Ga. 1976), Judge Freeman stated a more expansive approach. Relying on Carter v. Estelle, 519 F.2d 1136 (5th Cir. 1975), he said that upon an allegation of conspiracy under § 1985 between bank officials and state officials to instigate a false prosecution the bank which employs the involved officials could be liable vicariously.
. Section 1, Restatement (Second) of Agency (1958), states:
(1) Agency is the fiduciary relation which results from the manifestation of consent by one person to another that the other shall act on his behalf and subject to his control, and consent by the other so to act.
(2) The one for whom action is to be taken is the principal.
(3) The one who is to act is the agent. In serving as an exclusive hiring hall Local 542 was certainly acting on behalf of its members but was also acting on behalf of the employers in replacing traditional direct employment application processes. The bargaining agreement, either signed by the employers or otherwise adopted, manifests the requisite mutual consent. As to the right of control, the important issue is whether the employers retained power to oppose the union discrimination. As pointed out above, the employers were by contract entitled to a hiring hall operation which adhered to the terms provided. The contract required division of all applicants into experience-related classifications and referral in accordance with specified priorities. These mat*412ters were not beyond the potential control of the employers. Furthermore, a grievance procedure was established (for all relevant periods) by which complaints by applicants against the operation of the hiring hall could be made. This procedure would involve participation of an employer representative. Also, the bargaining agreement expressly provided since May, 1971, that referrals, the ultimate product of the hiring hall, not be made on a racially discriminatory basis. These indicia give far more by way of control than existed in United States Steel, supra, where not even express managerial resistance was effective to stop a master mechanic from choosing only union members as assistants and where that master mechanic was held to be an “agent” of the employer tying him to the illegal hiring. Agency principles establishing traditional bases for the spreading of risk and responsibility which the doctrine of respondeat superior seeks generally to accomplish, serve as an important guide in this case, just as they served to guide the court in Morrison-Knudsen and United States Steel.
. It should be noted also that comment a to § 214, Restatement (Second) of Agency, in defining liability for a breach of a non-delegable duty states:
[0]ne may have a duty to see that due care is used in the protection of another, a duty which is not satisfied by using care to delegate its performance to another but is satisfied if, and only if, the person to whom the work of protection is delegated is careful in giving the protection.
Duties arising under § 1981, manifestly, are no more subject to being bargained away than were the Title VII duties in Lorillard.
. Nonparties in January 1978 demanded a jury trial. This demand was made for four years after notice of this suit and months after the end of the lengthy year-and-one-half trial in this matter. Apart from the very serious question whether any nonparty member of a class, although permitted to bring its views to the attention of the court, can ever make such a demand without intervening under rule 24, F.R. Civ.P., the jury trial demand here must be foreclosed as untimely under rule 38(d), F.R.Civ.P. The demand was precipitated by a letter from plaintiffs announcing that contrary to their pri- or expectation they would not move to amend their complaint. This circumstance does not alter my view.
. The basis for liability of the class of defendant associations and contractors is § 1981. The redefinition of defendant class to comport with the six-year statute of limitations is therefore appropriate:
(a) All employer associations who have been parties to a collective bargaining agreement with defendant Local 542 at any time between November 8, 1965, and the date of this decree; and
(b) All employers who have been parties to a collective bargaining agreement with Local 542 and who hired at least one operating engineer affiliated with Local 542 at any time between November 8, 1965, and the date of this decree.
. Only Glasgow itself need be considered in evaluating the adequacy of representation. The presence of four associations as named defendants reassures that a thorough representation was provided.