Defendant Board of Examiners of the City of New York appeals from a pre*1169liminary injunction granted by Walter R. Mansfield, J.,1 in the United States District Court for the Southern District of New York at the instance of plaintiffs Boston M. Chance and Louis C. Mercado, suing on behalf of themselves and others similarly situated.2 The named plaintiffs, who are respectively black and Puerto Rican, sued the Board of Examiners under federal civil rights laws, 42 U.S.C. §§ 1981, 1983. Plaintiffs claimed that competitive examinations given by the Board to those seeking permanent appointment to supervisory positions in the City’s schools discriminated against blacks and Puerto Ricans and violated the Equal Protection Clause of the fourteenth amendment.3 Judge Mansfield, after making extensive findings of fact, found sufficient merit in plaintiffs’ case to preliminarily enjoin the Board from using the examinations. The judge wrote two opinions, one in July 1971, reported at 330 F.Supp. 203 (1971), and the other, an unpublished Memorandum, in September 1971. A preliminary injunction was entered shortly thereafter. In February 1972, we heard the appeal, after receiving a sizeable number of amicus briefs because of the sensitive nature of the issues raised, the effect of their resolution on the contending parties, and the understandable intensity with which positions have been taken. Despite the extensive — and in some instances emotional — briefing and argument, the ultimate issues before us are simple to state: Were Judge Mansfield’s findings of fact clearly erroneous, and did he commit any errors of law? We discuss the issues of fact, in Parts II and III of this opinion and the questions of law in Parts IV and V. Finding the answer to both basic questions in the negative, we affirm.
I
The appeal comes to us in an unusual posture. Since plaintiffs attacked the method used to fill supervisory positions in the school system of the City of New York, one would surmise that their primary opposition would come from those in charge of that system, the Board of Education of the City of New York and its Chancellor, Harvey B. Scribner,4 both named as defendants in this action. However, although the Board of Education appeared below, it did not actively oppose the motion for a preliminary injunction and has not appealed from the district court’s order. The Chancellor has done even less. In a memorandum to the Board of Education, quoted by Judge Mansfield in his opinion, Mr. Scribner stated that to defend against plaintiffs’ case
would require that I both violate my own professional beliefs and defend a system of personnel selection and promotion which I no longer believe to be workable.
330 F.Supp. at 219-220. The Board of Examiners, however, vigorously opposed plaintiffs’ motion in the trial court and has appealed. Supported by some of the amici,5 it offers a number of arguments *1170why the district court’s order should be reversed. Before examining them, we will briefly recite the facts as found by Judge Mansfield or as they appear in the record.
To obtain a permanent supervisory job in the New York City school system an applicant must not only meet State requirements but also obtain a City license. This dual qualification is in effect in New York State only in Buffalo and New York City; elsewhere in the State, certification by State authorities alone is enough. Moreover, only the New York City School District maintains a Board of Examiners. The Board of Examiners was established by the State legislature near the turn of the century as an independent body to conduct examinations to be used in selecting New York City school system professional personnel. The Board of Education and the Chancellor prescribe the minimum education and experience requirements for supervisors, but the Board of Examiners prepares and administers the examinations. The examination process itself may take as long as two years to complete. If a candidate successfully completes the examination, his name is placed on a list of those eligible for the particular supervisory post; he may then be selected by an appropriate school governing authority to fill an open position for which he is certified. If he is not appointed within four years after being placed on the list, his name is dropped and will not be re-listed until he again passes the examination.
For many hopefuls, the stumbling block to a permanent supervisory position has been the examination prepared by defendant Board of Examiners. That is true for plaintiffs Chance and Mercado, who have necessary State certificates and who meet the educational and experience requirements established by the City Board of Education. Chance has been employed in the New York City public school system for 15 years, Mercado for 12. They are now serving as acting principals of elementary schools in New York City, selected for those positions by their local community school boards. Unless they pass the Board’s examination, however, they are foreclosed from being appointed permanent principals. They brought this suit to challenge that obstacle as racially discriminatory and, therefore, unconstitutional.
Before reaching the merits of plaintiffs’ claim, the district court ordered the parties to develop a survey procedure to determine comparative pass rates of the different ethnic groups who had taken various supervisory examinations in recent years. The Survey was completed and is described more fully below. Despite the splendidly motivated genesis of the Board of Examiners,6 its examinations, according to the district court, have led to unintentional racial discrimination. After receiving extensive statistical evidence, Judge Mansfield found that:
[T]he examinations prepared and administered by the Board of Examiners for the licensing of supervisory personnel, such as Principals and Assistant Principals, have the de facto effect of discriminating significantly and *1171substantially against Black and Puerto Rican applicants.
330 F.Supp. at 223. The judge further found:
Such a discriminatory impact is constitutionally suspect and places the burden on the Board to show that the examinations can be justified as necessary to obtain Principals, Assistant Principals and supervisors possessing the skills and qualifications required for successful performance of the duties of these positions. The Board has failed to meet this burden.
Id. Because he believed that there was a strong likelihood that plaintiffs would ultimately prevail on the merits at trial and that the balance of the equities rested with plaintiffs, the judge enjoined any further examinations and licensing based upon previous examinations until “determination” of the action or until the Board of Examiners had satisfactorily revised its examination procedures.
II
Arguing that the injunction now in effect should be reversed, defendant Board of Examiners raises a number of issues on appeal, some primarily factual, others questions of law. In the former category, the Board claims that the statistics before the trial judge had little probative value and were used by him improperly and that the judge’s finding that the Board’s supervisory examinations were not job-related was clearly erroneous. In opposing plaintiffs’ motion for injunctive relief in the district court, the Board itself had argued that:
If statistics have any relevancy in determining the claim of discrimination, the only meaningful statistic would be a comparison of the pass-fail ratio of [whites with those of] black and Puerto Rican applicants.7
As already indicated, the court thereafter ordered the parties to develop a Survey to determine comparative pass rates of different ethnic groups in recent years. The Survey took several months to complete and covered 50 supervisory examinations given during the last seven years, involving approximately 6,000 applicants. After both sides submitted affidavits of experts and briefs, there was an evidentiary hearing directed to the relevance and significance of the tabulations. The parties were then given a further opportunity to present evidence on the issue of job-relatedness. The record shows that the trial judge in the early stages of the case was skeptical of plaintiffs’ ability to prove their claims, and thereafter proceeded cautiously, thoroughly and fairly before making any findings of fact. It is against that background that defendant’s attack on his factual findings must be considered.
The district judge read the Survey to show that white candidates passed the various supervisory examinations, considered together, “at almost 1% times the rate of Black and Puerto Rican candidates.” 330 F.Supp. at 210. The court, however, found even more significant the fact that:
[W]hite candidates passed the examination for Assistant Principal of Junior High School at almost double the rate of Black and Puerto Rican candidates, and passed the examinations for Assistant Principal of Day Elementary School at a rate one-third greater than Black and Puerto Rican candidates.
Id. The statistics for the latter two examinations were thought particularly significant “because they were taken by far more candidates than those taking any other examinations conducted in at least the last seven years,” and “because the assistant principalship has traditionally been the route to and prerequisite for the most important supervisory position, Principal.” Id. The judge reasoned that these examinations for assistant principal screened minority applicants out of a chance to become full principals, thus in effect magnifying the overall statistical differences between *1172white and non-white pass-fail rates. 330 F.Supp. at 210-211.8 The court also relied on other statistics, which showed that cities not using New York’s system of examinations had a startlingly higher percentage of blacks and Puerto Ricans in supervisory positions:
Detroit 281 16.7% 16.7%
Philadelphia 267 16.7% 16.7%
Los Angeles 1,012 8.0% 1.7% 9.7%
Chicago 479 6.9% 6.9%
New York 862 1.3% 0.1% 1.4%
Detroit 360 24.7% 0.2% 24.9%
Philadelphia 225 37.0% 37.0%
Los Angeles - -
Chicago 714 32.5% 32.5%
New York 1,610 7.0% 0.2% 7.2%
330 F.Supp. at 213. Thus, the percentages of New York City black and Puerto Rican principals and assistant principals (1.4% and 7.2%) were found to be substantially below those percentages for other large city school systems. The percentage of black and Puerto Rican principals in Detroit and Philadelphia (16.7%), for example, was 12 times higher than that in New York (1.4%).
Appellant Board offers a number of arguments challenging the court’s use of these statistics. Focussing on the Survey, the Board claims that the pass rates for the racial groupings of candidates were not comparable because the black and Puerto Rican group was a disproportionately large percentage of the eligibility pool for that group and, therefore, included more less competent candidates than did the white group of applicants. This was apparently not the position the Board urged below, nor was substantial evidence introduced bearing on the assumptions upon which the argument is based. Absent such evidence, we would not say that the trial judge could not regard the respective pass rates as probative of discrimination. The Board also argues that the results of examinations for many different jobs could not be lumped together for purposes of analysis. However, plaintiffs’ expert testified that this was quite proper and the judge was entitled to rely on that evidence. The Board also claims that the court erred in relying on comparative pass rates for all applicants instead of eliminating those who withdrew before completing the entire examination process. Here, too, the record contains expert evidence supporting the court’s method. The Board also argues that the statistics are meaningless because the samples were not random but were self-selected since examinations were taken by individual choice. As we have already indicated, the trial judge had sufficient evidence before him to use the Survey statistics as he did.
The Board further objects to the court’s findings that examinations for assistant principal in effect serve to screen out minority candidates from becoming principals and thus magnify the discriminatory impact of the examination system. This conclusion was based on the court’s understanding that “the assistant principalship has traditionally been the route to and prerequisite for the most important supervisory position, Principal.” 330 F.Supp. at 210. Relying on a September 1970 “Examination Announcement” for the November 1970 examination given for Principal of a Day Elementary School, the Board claims that the court misunderstood the requirements for becoming a principal.9 The Board also argues that the supposed cumulative discriminatory effect is *1173shown to be spurious by the pass-fail data on the 1970 elementary school principal examination. According to the Board, the pass-fail rate of black and Puerto Rican candidates on that examination was substantially higher than in the past and was comparable to the rate for the white group of candidates. As to the former argument, it is enough to point out that the trial court stated that the position of assistant principal was the traditional “route to and prerequisite for” becoming a principal. From our reading of the record, this appears to be an accurate statement for the period covered by the Survey. That the eligibility requirements for the 1970 examination were more liberal than in the past does not undermine the court’s reasoning. As to the relevance of the Board’s pass-fail statistics derived from the 1970 examination for principals, the data was produced after the district judge’s decision in July 1971, and was rejected by him in his September memorandum on the grounds that the information was incomplete and that plaintiffs had not had an opportunity to cross-examine or to gather all relevant facts regarding that examination.10
Finally, the Board claims that the comparison of the percentages of black and Puerto Rican principals and assistant principals in New York City with the percentages in the four other largest school systems in the country is invalid because until recently at least the New York City education and experience eligibility requirements were higher than those in other cities. Thus, the Board argues, the smaller percentage of black and Puerto Rican supervisors in New York City was due to those requirements rather than to the examinations. But, regardless of what light this may throw upon the validity of the education and experience requirements, it still does not follow that the comparison is wholly invalid. The inference was open to the trial judge on the record before him that the difference in the percentages in the large cities was due to the type of examination apparently given only in New York.
The last argument points up the crucial nature of the fact finding process in a case like this. After all the technical statistical jargon like “one tail” or “two tail” tests and “Chi-Square Test (Yates-corrected)” as well as the less esoteric numbers and percentages were placed before the trial judge, it was his job to resolve the issues. Throughout the briefs of the Board and its supporters runs the argument that other reasons can be inferred from the record for the comparatively low numbers of blacks and Puerto Ricans in supervisory positions. That may very well be true. But the question before us is whether the trial judge on the record before him was required to accept those inferences, and it is quite clear that he was not. In sum, while not all of us might have made the same factual inferences of racially discriminatory effect from the statistical evidence, both documentary and oral, before the court, none of us can say with the firm conviction required that those factual findings were mistaken. See United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 788, 92 L.Ed. 1147 (1948).
Ill
The parties agree that the case did not end with the district court’s finding that the examinations prepared and administered by the Board significantly and substantially discriminated against black and Puerto Rican applicants. The district court pointed out that “the existence of such discrimination, standing alone, would not necessarily entitle plaintiffs to relief.” 330 F.Supp. at 214. The further question was whether the exam*1174inations could be “validated as relevant to the requirements of the positions for which they are given, i. e., whether they are ‘job related.’” Id. On that issue the court first held that the Board had the burden of making a “strong showing” that the tests were in fact job-related; it then concluded that the burden was not met. The correctness of these rulings as well as the constitutional issues surrounding them are discussed in the following section of this opinion. For the moment, however, we confine our attention to the court’s factual finding that the examinations had not been validated.
The district court pointed out that two generally accepted methods are used to determine whether a particular examination is job-related or reasonably constructed to measure what it purports to measure. One is “content validation,” which requires the examiners to demonstrate that they have formulated examination questions and procedures based on an analysis of the job’s requirements, usually determined through empirical studies conducted by experts. An examination has content validity, then, if it elicits “from the candidate information that is relevant to the job for which it is given.” 330 F.Supp. at 216. The other method of evaluating job-relatedness is “predictive validation,” which requires a showing that there is a correlation between a candidate’s performance on the test and his actual performance on the job. Part of the controversy in the trial court was over which of these was the better method for evaluating the Board’s examinations. Plaintiffs stressed “predictive validation” as the more relevant test, while defendant argued that “content validation” was the more useful criterion. The district court did not resolve the dispute11 since it concluded that the Board had failed to establish that examinations were “valid as to content, much less to predictiveness.” 330 F.Supp. at 219.
In reaching its finding the court had to choose between conflicting expert testimony covering the issue of job-relatedness. Defendant Board submitted affidavits of several respected leaders in the field of educational testing, who stated that on the information supplied them the Board was apparently following testing methods that reasonably assured content valid examinations. The Board also introduced several research reports written by its staff members as showing its efforts to insure job-relatedness. Against this evidence plaintiffs offered the affidavits of various experts who found the Board's examinations lacking in validity, whether content or predictive. In making his finding the trial judge obviously relied heavily on the expert evidence offered by plaintiffs. The judge noted that the “fatal weakness in the Board’s system” lies in its failure to actually implement the “techniques and procedures adopted in principle and approved by independent experts.” Id. He further found that the Board’s research reports were either irrelevant to developing valid examinations or were grossly inadequate for that purpose. Finally, the court stated that its conclusion based on the expert testimony was “confirmed” by its own study of some of the examinations. The trial judge seems to have felt that, at least from a layman’s perspective, the examinations placed more emphasis on measuring a candidate’s ability to memorize than on his ability to perform as a supervisor. It should be pointed out, however, that the district court’s finding of invalidity was limited to the written part of the examinations. The judge made “no finding as to the content validity of the oral examinations, standing alone.” 330 F.Supp. at 222. The court also held that the evidence was insufficient to *1175support a finding that the Board had failed to administer the examinations objectively.
The Board argues that the district court’s finding that the written parts of the examinations were not shown to be sufficiently job-related is clearly erroneous. It claims that the court unreasonably rejected the opinions of the Board’s experts and instead ignored the “plain import” of their testimony, and improperly relied on irrelevant evidence and on the court’s own inexpert judgment. We are unpersuaded by these arguments. The justification for any written examination must at least be, as one of plaintiffs’ experts pointed out, that using it is better than drawing names out of a hat.12 Staying with that perhaps over simple idea for a moment, the use of a test should achieve much more than that. Judge Mansfield found that the tests at issue here did not. One cannot read the documents submitted by plaintiffs without experiencing great doubt over whether a lower score on the Board’s examinations necessarily meant poorer job qualifications or without wondering whether the examinations tested anything other than the ability to take a certain kind of test. The affidavits of plaintiffs’ main expert certainly suggest that familiarity with esoteric words, general cultural knowledge, expertise in current events and flawless English grammar might have far less to do with excellent job performance as a supervisor than other qualities not adequately tested.13 The judge did not hold or even imply that tests should be so constructed that blacks and Puerto Ricans would have to do well. But he did find that:
Although it has taken some steps towards securing content and predictive validity for the examinations and has been improving the examinations during the last two years, the Board has not in practice achieved the goals of constructing examination procedures that are truly job-related.
330 F.Supp. at 223. We cannot say that the judge erred. It is clear, of course, that he was not required to accept the views of the Board’s experts. In sum, what we said earlier applies here as well: While not all of us might have made the same factual finding on the question of job-relatedness as the district judge did, his finding was not clearly wrong.
IV
We come, then, to the question whether the district court applied the proper constitutional standards in reaching its conclusions that (1) plaintiffs made out a prima facie case of racial discrimination, and (2) the Board, in turn, failed to meet its burden of demonstrating that the examinations are justified notwithstanding- their discriminatory effect.
Concededly, this ease does not involve intentionally discriminatory legislation, cf. Loving v. Virginia, 388 U.S. 1, 87 S.Ct. 1817, 18 L.Ed.2d 1010 (1967), or even a neutral legislative scheme applied in an intentionally discriminatory manner, see Yick Wo v. Hopkins, 118 U.S. 356, 6 S.Ct. 1064, 30 L.Ed. 220 (1886). Nonetheless, we do not believe that the protection afforded racial minorities by the fourteenth amendment is exhausted by those two possibilities. As already indicated, the district court found that the Board’s examinations have a significant and substantial discriminatory impact on black and Puerto Rican applicants. That harsh racial impact, even if unintended, amounts to an invidious de facto classification that cannot be ignored or answered with a shrug. At the very least, the Constitution requires that state action spawning such a classification “be justified by legitimate *1176state considerations.” Abate v. Mundt, 403 U.S. 182, 185, 91 S.Ct. 1904, 1906, 29 L.Ed.2d 399 (1971). See Whitcomb v. Chavis, 403 U.S. 124, 149-160, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971); United States ex rel. Chestnut v. Criminal Ct. of the City of New York, 442 F.2d 611, 618 (2d Cir.), cert. denied, 404 U.S. 856, 92 S.Ct. 111, 30 L.Ed.2d 98 (1971); Kennedy Park Homes Ass’n v. City of Lackawanna, 436 F.2d 108, 114 (2d Cir. 1970), cert. denied, 401 U.S. 1010, 91 S.Ct. 1256, 28 L.Ed.2d 546 (1971); Norwalk CORE v. Norwalk Redevelopment Agency, 395 F.2d 920, 931 (2d Cir. 1968); Western Additional Community Organization v. Alioto, 340 F.Supp. 1351 (N.D.Cal. Feb. 7, 1972); Penn v. Stumpf, 308 F.Supp. 1238 (N.D.Cal. 1970); Arrington v. Massachusetts Bay Trans. Auth., 306 F.Supp. 1355 (D.Mass. 1969). In Kennedy Park Homes Ass’n, supra, Mr. Justice Clark, sitting on this court by designation, stated (436 F.2d at 114):
Even were we to accept the City’s allegation that any discrimination here resulted from thoughtlessness rather than a purposeful scheme, the City may not escape responsibility for placing its black citizens under a severe disadvantage which it cannot justify.
The Board argues, however, that the statistical differences found by the district court are insufficient to meet the constitutional test of invidiousness and do not amount to a prima facie case of de facto discrimination. In particular, the Board notes that even according to the district court’s analysis, on an overall basis white candidates passed at only one and one-half times the rate of black and Puerto Rican candidates. Such a difference, it is argued, is at most mere underrepresentation and hardly amounts to the gross unexplained disparity that is required for a prima facie case. See, e. g., Swain v. Alabama, 380 U.S. 202, 85 S.Ct. 824, 13 L.Ed.2d 759 (1965). But as we previously indicated, the district court did not merely rely on the difference in overall pass rates. Of “greater significance” were the pass rates with respect to the assistant principal examinations and the magnifying effect that results from requiring candidates to pass the examinations seriatim. Moreover, an additional factor considered was the small percentge in New York City of minority principals and assistant principals in comparison with other large metropolitan school systems that do not have comparable examination requirements. We believe that on this and other evidence in the record the district court could properly conclude that plaintiffs had demonstrated a disparity of sufficient magnitude to amount to a prima facie case of invidious de facto discrimination.
We further believe that once such a prima facie case was made, it was appropriate for the district court to shift to the Board a heavy burden of justifying its contested examinations by at least demonstrating that they were job-related. First, since the Board is specifically charged with the responsibility of designing those examinations,14 it certainly is in the better position to demonstrate their validity. Cf. Cooper and Sobel, Seniority and Testing Under Fair Employment Laws: A General Approach to Objective Criteria of Hiring and Promotion, 82 Harv.L.Rev. 1598, 1665 (1969). Second, once discrimination has been found it would be anomalous at best if a public employer could stand back and require racial minorities to prove that its employment tests were inadequate at a time when this nation is demanding that private employers in the same situation come forward and affirmatively demonstrate the validity of such tests. See Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e; Griggs v. Duke Power Co., 401 U.S. 424, 432, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971). The anomaly would only be emphasized by the recent passage of the Equal Employment Opportunity Act of 1972, which *1177broadened Title VII to include state and city public employers.15
The Board maintains, however, that the district court applied an improper standard in determining whether the examinations had been justified notwithstanding their discriminatory impact. According to the Board, the district court “clearly erred” in applying the compelling interest standard rather than the rational relationship test customarily applied in equal protection cases.
Although state action invidiously discriminating on the basis of race has long called for the “most rigid scrutiny,” Korematsu v. United States, 323 U.S. 214, 216, 65 S.Ct. 193, 89 L.Ed. 194 (1944), the Supreme Court has yet to apply that stringent test to a case such as this, in which the allegedly unconstitutional action unintentionally resulted in discriminatory effects. See Dandridge v. Williams, 397 U.S. 471, 485 n. 17, 90 S.Ct. 1153, 25 L.Ed.2d 491 (1970); cf. Whitcomb v. Chavis, supra, 403 U.S. at 149-160, 91 S.Ct. 1858, 29 L.Ed.2d 363. Manifestly, the question whether that test should be applied to de facto discriminatory classifications is a difficult one and is not to be resolved by facile reference to cases involving intentional racial classifications. We think, however, that the district court’s decision may be upheld under the “more lenient equal protection standard” and so find it unnecessary to reach this most difficult question. Eisenstadt v. Baird, 405 U.S. 438, 447 n. 7, 92 S.Ct. 1029, 1035, 31 L.Ed.2d 349 (1972); see Turner v. Fouche, 396 U.S. 346, 362, 90 S.Ct. 532, 24 L.Ed.2d 567 (1970).16 To .be sure, the district court stated that the de facto classification found was “constitutionally suspect” and that the Board was required to make a “strong showing” that its examinations can be “justified as necessary.” 330 F.Supp. at 216, 223. Such language usually connotes application of the “compelling interest” test. But the court’s actual analysis indicates that it never reached the point where application of that test would bring a different result from application of the rational relationship test. It is true that the court placed a heavy burden of proof on the Board— properly so, as we have already indicated. But the proposition to be proved was only that the Board’s examinations were job-related. As to that, the court concluded that the Board’s procedures for insuring the basic content validity of its tests were inadequately implemented and that as a result the tests themselves did not measure what they purported to measure. In short, the present examinations were not found to be job-related and thus are “wholly irrelevant to the achievement of a valid state objective.” Turner v. Fouche, supra, 396 U.S. at 362, 90 S.Ct. at 541; see Reed v. Reed, 404 U.S. 71, 75-76, 92 S.Ct. 251, 30 L.Ed.2d 225 (1971). The court did not reach the issue whether — or even suggest that — if the written examinations were job-related the Board would still be required to demonstrate that no less discriminatory means of obtaining its supervisory personnel were available. See Loving v. Virginia, supra, 388 U.S. at 11, 87 S.Ct. 1817; McLaughlin v. Florida, 379 U.S. 184, 196, 85 S.Ct. 283, 13 L.Ed.2d 222 (1964); cf. United States v. Bethlehem Steel Corp., 446 F.2d 652, 662 (2d Cir. 1971). Had the court done that, the bite of the “compelling interest” test would apply. Instead the court apparently thought that appropriate examinations could be prepared,17 but that the Board *1178had not yet done so. The Board, then, failed to meet its burden even under the rational relationship standard, which would be the least justification that the Constitution requires.
V
The last issue before us is whether the district judge abused his discretion in issuing a preliminary injunction against the use of Board examinations and lists of eligibles based upon the results of those tests and in requiring the Board to allow appointment of “acting” supervisors. The judge concluded that there was a strong likelihood that plaintiffs would prevail on the merits at trial, a view we have already indicated we would not characterize as wrong. But that did not end the matter, as the judge recognized, because the impact of the interim order is sizeable. It temporarily prevents the use of outstanding lists of eligible licensed supervisory personnel and bars the issuance of new lists from recently completed examinations. We are dealing here not with abstractions but with people, some of whom have worked diligently for years to pass the Board examinations and rise to a more responsible job. If we needed any proof of that, it was furnished by the eloquent amicus brief and oral argument of one such teacher, Charles Wiener.18 Also, the injunction intrudes upon the operation of a school system in which clearly the local, rather than the federal, authorities, have primary responsibility. See Epperson v. Arkansas, 393 U.S. 97, 104, 89 S.Ct. 266, 21 L.Ed.2d 228 (1968).
But the district judge was aware of these and other consequences of his order. He pointed out that those now on the lists would not be denied “an equal opportunity in the future to qualify under such examination procedures as are found to be constitutionally permissible” and meanwhile would still be eligible for appointments in “acting” capacities. 330 F.Supp. at 224. Based upon his findings, he also recognized that refusing the injunction would continue racial discrimination and deny plaintiffs and others like them any real opportunity for appointment until a final decision on the merits, by which time many positions would have been filled. Moreover, many minority group acting supervisors, already selected by local boards because of their ability to perform, might lose their positions.19 Even more significantly, the court found that granting the injunction would cause no “great harm” to the public or to school children, in view of the availability of acting appointments, by no means unusual, see Council of Supervisory Ass’ns of Public Schools of New York City v. Bd. of Edue., 23 N.Y.2d 458, 297 N.Y.S. 2d 547, 245 N.E.2d 204 (1969), the possibility that new and better examination procedures would be devised, and the Chancellor’s declaration that the system then in effect was not “workable.” Taking all of these considerations into account, the court concluded that “the balance of hardship tips decidedly in favor of plaintiffs,” 330 F.Supp. at 224, thus applying the traditional test. See e. g., Checker Motors Corp. v. Chrysler Corp., 405 F.2d 319, 323 (2d Cir.), cert. denied, 394 U.S. 999, 89 S.Ct. 1595, 22 L.Ed.2d 777 (1969). This was typical of the hard — perhaps even agonizing — decisions that a trial judge must often make. But make them he must one way or the other since he is on the firing line. Our job is to decide whether Judge Mansfield’s decision to enjoin use of the tests, which was not unprecedented,20 was an abuse of discretion. We cannot say that it was.
We understand the fears of those individuals and groups that have filed *1179strong and even passionate briefs, urging us to reverse. We share their concern for the public school system; its strength is crucial in our society. But emotion has led some of the amici astray in describing the decision of the court below. The judge did not approve of a quota system for the appointment of supervisory personnel; he specifically rejected the idea.21 Nor did he permanently do away with the merit system and substitute nepotism and patronage. The judge did not outlaw other written examinations or indicate that none could be created to test more fairly the qualities necessary for a supervisory job.22 It may well be that new testing procedures will be devised by the parties themselves and be approved by the district court.23 Certainly, the case should not linger on in its present unfinished state. All that the court below did was to enjoin, on an interim basis, examinations that it justifiably found to have a discriminatory effect and to be ill-suited for their purpose and to allow the Board of Education of New York City and its Chancellor to fill vacancies on an acting basis with candidates meeting criteria satisfactory to them. That order was not improper and we affirm it.
We cannot close this opinion without a word of appreciation to the lawyers for the parties who have labored long and well to present the issues to us. They are listed at the beginning of this opinion, and they have our appreciation.
Judgment affirmed.
. Judge Mansfield heard the ease as a district judge and, after appointment to the circuit court, was designated to sit on the district court to complete this phase of the case.
. Six other named plaintiffs were added in an amended complaint filed in February 1971. At the time of Judge Mansfield’s decision in July 1971, decision was reserved on plaintiffs’ motion to have the action designated as a class action under Fed.R.Civ.P. 23.
. Plaintiffs also claimed violations of Art. 5, § 6 of the New York State Constitution (McKinney 1969) and §§ 2590-j (3) (a) (i), 2569(1), 2573(10) of the New York Education Raw (McKinney’s Consol. Raw c. 16, 1970). The district court did not pass on these claims and they are not at issue before us.
. Mr. Scribner is also the Chief Administrator of the School District of the City of New York.
. The following have submitted amicus briefs in opposition to the district court’s order: United Federation of Teachers, Local 2, American Federation of Teachers, AFR-CIO; Council of Supervisors and Administrators of the City of New York, Local 1, School Administrators and *1170Supervisors Organizing Committee, AFL-CIO; Anti-Defamation League of B’nai Brith; Council of Jewish Organizations in Civil Service, Inc.; Catholic Teachers Association of the Diocese of Brooklyn ; Oscar Gottlieb; and Charles Wiener. Amici in support of the district court decision are: Public Education Association ; Aspira of America, Inc.; and New York Association of Black Educators.
. The Board was
designed to do away with the abuses, real or supposed, arising from the appointment and promotion of teachers in the several parts of the city on a basis of social and religious favoritism and of political patronage, and to place the appointment and promotion of teachers purely on a competitive basis of merit.
Affidavit of defendant Examiner Gertrude E. Unser, dated Oct. 26, 1970, quoting 1923 statement of Superintendent of Schools, William Ettinger.
. Affidavit of defendant Examiner Jay E. Greene, dated Oct. 26, 1970.
. The judge also pointed to other statistical evidence found in the Survey; e. g., on six examinations taken by a minimum of 18 candidates from the white and nonwhite groups, the percentages of the former passing were from 9% to 28% higher than those of the latter. 330 F.Supp. at 213.
. Tlie Board claims that in 1970, the “eligibility requirements for examination and licensing as an elementary school principal require only one year’s experience in any supervisory position (acting or regular) or a college-supervised internship.” Appellant’s brief at 19.
. We note, without indicating a view, that the fact that the 1970 examination was changed in material respects, combined with the significant increase (asserted by appellant) in the pass rates of blacks and Puerto Ricans, arguably supports, rather than weakens, plaintiffs’ claim that the prior examinations had a substantial and significant racially discriminatory effect.
. The court did note, however, that:
Predictive validity is of greater significance in evaluating aptitude tests than proficiency tests. Furthermore it often takes a long time to establish such validity and even then the evaluation depends upon the reliability and fairness of the field appraisal of performance on the job.
330 F.Supp. at 216.
. Affidavit of Richard S. Barrett, dated Nov. 5, 1970.
. We are aware, as appellant points out, that these objections might not apply witii equal force to the 1970 examination for principal. But for reasons given above, see text at notes 9 and 10 supra, we do not regard that as undermining the court’s finding.
. See N.Y.Educ.Law §§ 2569(1), 2580-j (3) (a) (1) (McKinney 1970).
. We put to one side questions as to how this amendment affects the future course of this case in the trial court.
. We also need not decide whether, upon a showing of invidious discriminatory impact in a case such as this, the State is required to establish predictive validity as well as content validity for its job testing procedures.
. The preliminary injunction itself provided that any party “may ... apply for modification of this order to permit further examinations for supervisory positions, and the promulgation of lists, issuance of licenses, and making appointments based thereon. . . . ”
. The brief was also filed on behalf of Oscar Gottlieb, another teacher.
. See N.Y.Edue.Law § 2573(2) (McKinney 1970).
. See Western Additional Community Organization v. Alioto, 340 F.Supp. 1351 (N.D.Cal. Feb. 7, 1972); Hicks v. Crown Zellerbach Corp., 319 F.Supp. 314, 321 (E.D.La.1970).
. The court said (330 F.Supp. at 214):
The Constitution does not require that minority group candidates be licensed as supervisors in the same proportion as white candidates.
. See note 17 supra and accompanying text.
. Attached to plaintiff’s brief as “Appendix A” and “Appendix B” are documents containing proposals for modifying the Board’s examination procedures. These proposals are the result of the parties’ compliance with Judge Mansfield’s urgent request that they cooperate fully to devise a mutually acceptable method of selecting supervisors. However, the proposals were not part of the record in the district court. We have not considered them in reaching our decision, and we grant defendant Board of Examiners’ motion to strike appendices “A” and “B” of plaintiffs’ brief. We have also not considered the “Regulations Governing the Assignment of Acting Supervisors,” submitted to us by plaintiffs' after oral argument.