OPINION OF THE COURT
GIBBONS, Circuit Judge.
In this action, brought on March 15, 1972 by the Attorney General on behalf of the United States1 for relief from alleged violations of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq., the defendant International Union of Elevator Constructors, Local Union No. 5 (Local 5) appeals from a final order of the district court adjudging Local 5 in violation of the Act and granting injunctive relief that includes affirmative action to overcome the effects of the violation.2 This appeal challenges both the evidentiary support for the finding of violation and the scope of relief. We affirm.
I. THE VIOLATION
A. Background facts
Local 5 is a labor organization within the meaning of Title VII. 42 U.S.C. § 2000e(d). It is a member of the International Union of Elevator Constructors, and that parent union has a collective bargaining agreement with the National Elevator Manufacturing Industry, Inc., a national association of elevator construction contractors. The contract binds those contractors who operate in the Philadelphia area. The jurisdiction of Local 5 covers Bucks, Chester, Delaware, Montgomery and Philadelphia Counties in Pennsylvania; New Castle County, Delaware; and Atlantic, Burlington, Camden, Cape May, Cumberland, Gloucester, Mercer, Monmouth, Ocean and Salem Counties in New Jersey. Under the terms of the national collective bargaining agreement which was in effect for times here relevant, Local 5 was the exclusive bargaining representative for persons employed in the elevator construction industry in the area of its jurisdiction. That agreement classified union members as either “mechanics” or “helpers”. Helpers receive 70% of a mechanic’s wage. A helper must have completed two years in the industry and passed an examination administered by a joint employer-union committee to be eligible for mechanic status. To be eligible for union membership a worker must be 18 years of age and must have worked in the industry as a probationary helper for at least 100 hours a month for six months within a nine month period. Probationary helpers receive 50% of a mechanic’s wage, and receive increases automatically to 70% when the probationary period has been completed. A helper who has completed his probationary period is normally admitted to the union, and thereby made a “card helper” at an initiation scheduled by the union’s executive board. Initiations are usually held every 12 to 15 months, depending to some extent on employment trends in the industry.
Employment in the elevator construction trade is obtained, as a practical matter, through a hiring hall controlled by Local 5. While the collective bargaining agreement does not require that mechanics and helpers employed by contractors be members of the [1014]*1014union, it does require that those hirees who are not members obtain union referral cards. Thus any probationary helper must hold a work permit. The collective bargaining agreement compels employers to furlough probationary helpers before any others. It does not require that helpers who are not yet union members be laid off first, but it is customary for Local 5 to importune employers to furlough people on a reverse seniority basis and to lay off non-member helpers before card helpers.
Under the 1967 agreement, Local 5 followed a policy of denying work permits to non-members while members were out of work. This informal practice was formalized in Article XXII of the 1972 collective bargaining agreement. The hiring preference extends not only to Local 5 members who are unemployed, but to members of any other local of the International Union who want their names included on the Local 5 employment list.
In 1969 the Secretary of Labor, acting pursuant to executive order, implemented the so-called “Philadelphia Plan” which required bidders on any federal or federally assisted projects in a five-county area in and around Philadelphia to submit affirmative action plans, and which imposed a 19%-23% minority manpower goal for that trade. The plan was upheld by this court against charges that it was inconsistent with the Civil Rights Act of 19643 and the National Labor Relations Act,4 and that it was unconstitutional. Contractors Association v. Secretary of Labor, 442 F.2d 159 (3d Cir.), cert. denied, 404 U.S. 854, 92 S.Ct. 98, 30 L.Ed.2d 95 (1971). That opinion described the prior history of underrepresentation of blacks and other minorities in the construction trades in the Philadelphia area. We held that the Secretary of Labor could, because of the federal interest in maximum availability of construction tradesmen for projects in which the government had a cost and completion interest, require affirmative action without a finding as to the reason for exclusion of available tradesmen from the labor pool. 442 F.2d at 177.
B. The government’s case
The Attorney General’s proposed remedy in this case is broader in scope than the Secretary’s Philadelphia Plan, for it is aimed at all projects, not merely federal projects, and at all of Local 5’s jurisdiction. The complaint charges a longstanding policy or practice of discrimination against blacks, and seeks broad injunctive relief. It also charges, however, that Local 5 has interfered with and prevented the implementation of the Philadelphia Plan. Thus the complaint seeks injunctive relief both to implement Executive Order No. 11246 and the Plan which Contractors Association v. Secretary of Labor sustained, and to enforce Title VII. The district court’s opinion and judgment, however, predicate relief only upon the finding of a Title VII violation. The court made no specific findings relating the prescribed relief to implementation of the Philadelphia Plan. Thus unlike the Philadelphia Plan case, we may not affirm on the narrow ground that proof of underrepresentation of black tradesmen in the labor pool is a sufficient basis for relief. There must be proof of a pattern or practice of discrimination. 42 U.S.C. § 2000e-6(a).
At trial the government produced statistical evidence tending to show that blacks were grossly underrepresented in the membership of Local 5. At the time the suit was filed black membership was less than 1%. During the year and a half the suit was pending this figure rose to 3%. In contrast, the government offered census data5 from the 1970 Census which established:
[1015]*1015
City of Philadelphia City and Pennsylvania Suburbs of Philadelphia Local 5’s Jurisdiction
Percent of blacks in the total population 33.5% 19.4% 15.9%
Percent of blacks in the civilian labor force of males over 16 28.5% 15.9% 13.0%
Percent of blacks in the unemployed civilian labor force of males over 16 43.3% 31.6% 26.1%
Percent of blacks in the category of unemployed males over 16 whose last occupation was “Craftsman, Foreman, or Kindred worker.” 36.5% 24.0% 20.1%
When this suit was commenced in the Spring of 1972, only 11 work permits had ever been issued by Local 5 to black men. Two of these were issued pursuant to court orders following investigations and complaints by the Philadelphia Commission on Human Relations. Between 1967 and 1972, 215 work permits (the only potential means of gaining entry to the trade) were issued, of which 9 (4%) went to blacks. In the year and a half after the suit was filed, however, blacks received 51 permits, or 21% of the total. The evidence also showed that prior to October 1973, Local 5 referred a black union member to an employer who requested help on only one occasion, and that the union never has referred a black probationary helper in response to an employer request. It is true that between May 1967 and March 1972 when the complaint was filed, the issuance rate of work permits to black applicants (9 of 17, or 53%) rate was higher than the overall issuance rate (49%) during that period. But the issuance rate must be considered in the context of the vast disproportion in numbers between white and black applicants.
The government also offered evidence of alleged overt discrimination in seven specific instances. The district court found only two of the specific allegations supported by the evidence. On these two occasions blacks were denied work permits, were told that some union members and permit holders were out of work, and were told that policy prevented the issuance of permits to new workers under these circumstances. The court found, however, that permits were issued to white applicants during the same period.
Local 5 urges that the district court erred in holding that the government’s evidence sufficed to establish a pattern or practice of discrimination. It does not contend that the court erred in finding that two specific instances of overt discrimination occurred. Rather, it urges that the government’s statistical presentation, even when bolstered by the inferences which might be drawn from those specific instances, did not establish a prima facie case of a pattern or practice of discrimination. Conceding that statistical evidence has repeatedly been held sufficient to establish a prima facie Title VII violation,6 it neverthe[1016]*1016less urges that the government’s statistical presentation in this case was deficient. It contends, first, that the 1970 census data was unreliable because it included persons 16 and 17, while the trade customarily employs only males age 18 and older. But it seems unlikely that the percentage of 18 year-old blacks in the work force would differ markedly from the percentage of 16 year-olds. In any event, the suit was filed in 1972, and by that year persons who were 16 in 1970 would be 18 and hence eligible for employment in the trade.7 This objection clearly lacks merit.
The union also complains that the government improperly relied on the census figure showing that 43% of the unemployed civilian labor force over age 16 was black as establishing the population figure against which the union’s minority enrollment should be measured. Local 5 argues that this demographic figure is misleading, because neither the heavily black City of Philadelphia nor the disproportionately black unemployed civilian work force over age 16 is an appropriate reference population. We are inclined to agree that the most significant piece of demographic data for purposes of this case is the percentage of black males over age 16 in the available and eligible work force. Nevertheless, the evidence tended to show that at least half of the elevator constructors in Local 5’s jurisdiction were employed at job sites in Philadelphia, and between 25% and 30% worked in the four counties surrounding that city. The remaining workers, perhaps as few as 15%-20%, worked in outlying counties in New Jersey and Delaware, where white populations largely predominated. Weighting the census data account for the distribution of employment opportunities, it would appear that when this suit was filed, blacks held 1% of the jobs in the elevator construction trade in areas where the available eligible work force was about 20% black. In such circumstances we are entirely satisfied by the strength of the government’s statistical showing. Many courts have found or sustained findings of violations on less impressive data. See, e. g., United States v. Ironworkers, Local 86, supra (.1% black union membership; 7% black population); Parham v. Southwestern Bell Telephone Co., supra (2% black employment; 22% black population); Rios v. Steamfitters, Local 688, 501 F.2d 622 (2d Cir. 1974) (1% non-white union membership; 20% nonwhite work force); Fowler v. Schwarzwalder, 351 F.Supp. 721 (D.Minn.1972) (1.3-2.2% minority employment; 4.6-8% minority population); Dozier v. Chupka, 395 F.Supp. 836 (S.D.Ohio 1975) (2.3% black employment; 9.9-18.5% black population); Sagers v. Yellow Freight System, Inc., 388 F.Supp. 507 (N.D.Ga.1973) (0-3.6% black employment; 10% black work force); Boston Chapter, NAACP, Inc. v. Beecher, 371 F.Supp. 507 (D.Mass.), aff’d, 504 F.2d 1017 (1st Cir. 1974), cert. denied, 421 U.S. 910, 95 S.Ct. 1561, 43 L.Ed.2d 775 (1975) (.2% minority employment; 13% minority population).
Moreover, the district court found, and the evidence shows, that about 75% of the persons applying to Local 5 for work permits were friends or relatives of union members. Fewer than 4% of the applicants were black. Several courts have held that word-of-mouth hiring practices that carry forward racial imbalances are themselves discriminatory. See Barnett v. W. T. Grant Co., 518 F.2d 543, 549 (4th Cir. 1975); Franks v. Bowman Transportation Co., 495 F.2d 398, 419 (5th Cir. 1975), rev’d on other grounds, 424 U.S. 747, 96 S.Ct. 1251, 47 L.Ed.2d 444, (1976); Parham v. Southwestern Bell Telephone Co., supra, 433 F.2d at 427. We simply cannot say, looking at the totality of the government’s evidence, that the district court erred in holding that a prima facie case of a Title VII pattern or practice violation was presented.
Local 5 next urges that the district court erred in dismissing as inadequate its evidence offered in rebuttal to the government’s case. Two lines of defense were [1017]*1017advanced. The first was that the elevator construction employers actually control employment opportunities. The second defense was that the economic recession that beset the construction industry during the years in issue made the recruitment of minorities infeasible. In evaluating these defenses it must be kept in mind that once a prima facie case of a Title VII violation has been proved, the burden shifts to the defendant to prove by a preponderance of the evidence that there is a benign explanation or justification.8
The district court accepted as true evidence that 65-75% of the elevator constructors were hired initially by employers. As explained above, the collective bargaining agreement permits employers to hire, and then to send the prospective employees to the Local for a work permit. No doubt employers do bear part of the responsibility for the historical underrepresentation of blacks in the industry labor force. See Contractors Association v. Secretary of Labor, supra. But in the remaining 25-35% of the cases the employer asks Local 5 for a referral. The district court found that Local 5 had never referred a black to an employer as a new probationary helper. The court concluded that
“whatever may be said of the employers’ recruitment practices, the uncontradieted evidence discloses that Local 5 has exercised its partial control of work opportunities in the trade to the complete exclusion of blacks.”
398 F.Supp. at 1255-56.
Local 5 points to the fact that from 1969 through 1973 its rate of issuance of permits to blacks exceeded the rate of issuance to whites. But as we pointed out above most of the blacks received permits only after this suit was filed. From 1969 to the filing of the suit 9 blacks and 215 whites received permits. We cannot say that the district court erred m concluding that evidence regarding the post-suit issuance of permits did not suffice to meet the union’s burden of explaining or justifying the prima facie violation prior to suit.
Finally, the union offered evidence that an economic recession in the construction industry made recruitment of black union members impracticable. For the fifteen month period following January 1, 1971, Local 5 issued only six work permits, in contrast with 107 during 1970. But while the downturn in construction during 1971 might bear on the practical ability of the union to take steps to correct the effects of past discrimination, it does not explain the fact that in 1969 permits were issued to 53 whites and no blacks, while in 1970 permits were issued to 101 whites and 6 blacks. The recession evidence simply did not meet the prima facie case, and was properly disregarded on the issue of the Title VII violation.
We conclude that the district court, in finding a Title VII pattern or practice violation, applied appropriate legal and evidentiary standards, and that the finding, to the extent that it is factual, is not clearly erroneous.
II. THE REMEDY
In fashioning a remedy for Local 5’s pattern or practice of discrimination, the district court granted four kinds of injunctive relief.9 It established a 23% black membership goal; it ordered a modification of the collective bargaining agreement to eliminate the hiring preference for members of foreign locals over Local 5 probationary helpers; it ordered a modification of the union seniority policies to prevent the Local from attempting to persuade employers to lay off workers in reverse order of seniority; and it ordered Local 5 during the life of the decree to adopt a 33% black [1018]*1018referral quota. The decree, unless modified, will last until the membership goal has been attained and maintained for one year. The union challenges each feature on numerous grounds. Many of the challenges go to the scope of equitable relief appropriate for the violation which was found, and are matters as to which the district court was required to exercise discretion. None, in our view, involved an abuse of the broad discretion in fashioning remedies with which a Title VII court is .vested. See Franks v. Bowman Transportation Co., supra, at 762-766, 96 S.Ct. 1251, 44 U.S.L.W. at 4360-61. They require no elaboration. Two legal contentions, however, involving both constitutional and statutory issues, merit discussion.
The constitutional contention is that both the membership goal and the referral quota are in themselves violations of due process or equal protection even when employed to overcome the effects of past discrimination. That contention is foreclosed by the settled law of this circuit. See Erie Human Relations v. Tullio, 493 F.2d 371 (3d Cir. 1974) and cases cited therein. See also United States v. T.I.M.E.-D.C., Inc., 517 F.2d 299 (5th Cir. 1975), cert. granted, - U.S. -, 96 S.Ct. 2200, 48 L.Ed.2d 814, (1976).
Nevertheless, Local 5 urges that, assuming the membership goal and referral quota can constitutionally be imposed as remedies, those remedies have been foreclosed to the federal courts by congressional action. Such a legislative design is said to be found in §§ 703(h) and 703(j) of Title VII, 42 U.S.C. §§ 2000e-2(h), (j),10 which the Local would have us interpret to prohibit both quota remedies and judicial interference with seniority systems. Local 5 misreads the statute.
The starting point for an analysis of the scope of the court’s remedial powers under Title VII is not § 703, which defines unlawful practices, but § 706,42 U.S.C. § 2000e-5, which contains the Title’s enforcement provisions. Section 706(g) provides:
If the court finds that the respondent has intentionally engaged in . . .an unlawful employment practice charged in the complaint, the court may enjoin the respondent from engaging in such unlawful employment practice, and order such affirmative action as may be appropriate, which may include, but is not limited to, reinstatement or hiring of employees, ... .or any other equitable relief as the court deems appropriate. [1019]*1019Nowhere in the enforcement provisions of Title VII is there any prohibition against the imposition of a membership goal or a referral quota to help achieve that goal. Indeed, both remedies would seem to be authorized by the broad language of § 706(g) empowering courts to order “such affirmative action as may be appropriate Cf. Franks v. Bowman Transportation Co., supra, at 768, 96 S.Ct. at 1264.
Local 5 contends, however, that the enforcement provisions in § 706(g) are in effect limited by the unlawful employment practices prohibitions in §§ 703(h) and (j). The short answer to that contention is that § 703 defines violations, not remedies. The section binds employers and labor organizations. It binds the court to the extent that it curtails the court’s power to find a violation, but § 703 is simply inapplicable to the relief which may be afforded once a violation has been found. Thus a court could not predicate a finding of violation upon the existence of a bona fide seniority system such as § 703(h) describes, or upon the mere failure to institute an affirmative action program of the kind to which § 703 (j) refers. But the court’s remedial powers are not limited by these provisions. That is made abundantly clear by the holding in Franks v. Bowman Transportation Co., supra, that artificial seniority may, pursuant to § 706(g), be awarded as a remedy for past discrimination. That case is dispositive of the Local’s contention that the remedy here interfered with a genuine seniority system, even assuming that the informal preference for union members which was in effect until 1972, and the formal arrangement embodied in the 1972 collective bargaining agreement can be so characterized.
No Supreme Court case has dealt specifically with a quota remedy under Title VII. But there is virtual unanimity in the lower federal courts that § 703(j) does not in this respect limit the courts’ § 706(g) remedial powers. Prior to 1972 at least four circuits rejected the “no quota remedy” contention, in the belief that “[a]ny [such] interpretation would allow complete nullification of the stated purposes of the Civil Rights Act of 1964.” United States v. IBEW, Local 38, 428 F.2d 144, 149-59 (6th Cir.), cert. denied, 400 U.S. 943, 91 S.Ct. 245, 27 L.Ed.2d 248 (1970). Accord, United States v. Ironworkers, Local 86, supra; United States v. Sheet Metal Workers Local 36, 416 F.2d 123 (8th Cir. 1969); Local 53, Asbestos Workers v. Vogler, 407 F.2d 1047 (5th Cir. 1969).
In 1972 Congress amended Title VII.11 The 1972 amendments did not make any significant change in the sections with which we are concerned, but a section-by-section analysis of the final version of H.R. 1746, the amending bill provides:
“In any area where the new law does not address itself, or in any areas where a specific contrary intention is not indicated, it was assumed that the present case law as developed by the courts would continue to govern the applicability and construction of Title VII.”
Subcomm. on Labor of the Senate Committee on Labor and Public Welfare, Legislative History of the Equal Employment Opportunity Act of 1972 at 1844 (1972). Congress was well aware of the case law referred to above. Senator Ervin was of the view that § 703(j) forbade the courts to use quota remedies to overcome the effects of past discrimination. Recognizing that both the EEOC and the courts had accepted the plain meaning of § 706(g), he proposed two amendments to S. 2515, the Senate equivalent of H.R. 1746, which would have restored his original understanding. Senator Javits, the principal spokesman against the Ervin amendments, specifically defended the pro-quota results in United States v. Ironworkers, Local 86, supra, and Contractors Association v. Secretary of Labor, supra, and had both opinions printed verbatim in the Congressional Record. The Senate rejected both Ervin amendments by two-to-one margins. See Legislative History, supra, at 1017,1042-74; 1681,1714-17. Thus we have in this instance unusually clear [1020]*1020evidence that Congress approved the pre1972 federal court interpretation of the scope of § 706(g) remedial power conferred by the 1964 Act. That indication is entitled to great weight. See Runyon v. McCrary, - U.S. -, -, 96 S.Ct. 2586, 48 L.Ed. -, (1976). Since 1972 the federal courts have naturally been quite firm in the view that quota remedies are permissible. See EEOC v. Local 638, 532 F.2d 821 (2d Cir. 1976); EEOC v. Detroit Edison Co., 515 F.2d 301 (6th Cir. 1975), petition for cert. filed, 44 U.S.L.W. 3215 (U.S.Sept. 12, 1975) (No. 75-393); Boston Chapter NAACP, Inc. v. Beecher, 504 F.2d 1017 (1st Cir. 1974), cert. denied, 421 U.S. 910, 95 S.Ct. 1561, 43 L.Ed.2d 775 (1975); Rios v. Steamfitters, Local 638, 501 F.2d 622 (2d Cir. 1974); NAACP v. Allen, 493 F.2d 614 (5th Cir. 1974); United States v. N. L. Industries, Inc., 479 F.2d 354 (8th Cir. 1973); Southern Illinois Builders Association v. Ogilvie, 471 F.2d 680 (7th Cir. 1972). We join those holdings.
Local 5 places great reliance upon this court’s decision in Jersey Central Power & Light Co. v. Local 327, IBEW, 508 F.2d 687 (3d Cir. 1975), vacated, -U.S. -, 96 S.Ct. 2196, 48 L.Ed.2d 812, (1976). We express no view whether or to what extent our holding in Jersey Central Power has been affected by these subsequent developments, for in any event that case is not in point because it did not deal with the scope of the judicial remedy afforded in § 706(g) for a § 703 violation.
The district court properly concluded that it could, under the authority vested in it by § 706(g), grant the four kinds of injunctive relief utilized in this case. The scope of relief fell well within the permissible range of district court discretion.
CONCLUSION
The court properly found that the union had committed a pattern or practice of racial discrimination in violation of Title VII. The court’s remedy was an appropriate response to that violation. The judgment appealed from will be affirmed.
This percentage does not include Cape May, Monmouth, and Ocean Counties. Figures are not available for experienced unemployed blacks in those counties.