CLARK, Associate Justice:I.

These consolidated cases are here on appeal from an order of the District Court denying injunctive relief, reinstatement and back pay to certain black teachers formerly employed by the County School Board of Nansemond County, Virginia. 351 F. Supp. 196 (1972). No. 73-1493 was originally filed on May 27, 1970 by the United States because of its dissatisfaction with the degree of desegregation in the Nansemond County School District under a freedom-of-choice plan. The action sought the total desegregation of the schools and faculties in the district controlled by the Board. Subsequently a desegregation plan submitted by the Board was ordered implemented by the Court for the school year 1970-71. The United States filed a motion in the case on June 1, 1971, for supplemental relief, alleging that the Board's plan had failed to disestablish the dual school system which had existed for many years in the district. It was further alleged that the Board had instituted a qualification and selection plan for the hiring and the retention of faculty members which had, in fact, substantially reduced its black teaching staff and that the Board had refused to demonstrate that its actions were not racially discriminatory. No. 73-1492 was filed against the Board on August 20, 1971, by thirteen black teachers as individuals and on behalf of the class represented by them. They alleged that the Board had racially discriminated against them and other black faculty members in violation of the Due Process and Equal Protection Clauses of the 14th Amendment by its application of certain testing criteria to them which resulted in the termination of their employment.

Prior to the 1970-71 school year the Board had required each applicant to have a baccalaureate degree, a Virginia teacher's certificate endorsed for the grade or subject to be taught and three letters of recommendation. The controversy here centers on a new requirement of the Board adopted January 13, 1970 that teachers take the National Teacher Examinations (NTE) and present a minimum score of 500 on the Weighted Common section of the examination. Teachers in non-academic areas were exempted, such as physical and driver education and certain trade and industrial courses. Currently employed teachers were permitted to continue in their positions without taking the NTE; however if a school transfer was effectuated, the teacher could be placed on a three-month provisional contract until such time as there was an opportunity to give the examination. The NTE is given several times a year by the Educational Testing Service of Princeton, New Jersey. It consists of a Weighted Common section and a Teaching Area section. The former offers a general appraisal of the teacher's basic professional preparation and general academic attainments upon graduation from college and before entering the teaching profession. The latter is a test of the individual's command of a specific academic subject. The Educational Testing Service does not recommend that the test be used as the sole criterion for employment; however the Board did so use it here, making it the sine qua non of employment.

The Board's action required that the test be taken by all teachers who came to the School District in the 1970-71 school year and all subsequent applicants. However, as implemented, only those teachers completing their first year of teaching with the School District were required to submit an NTE weighted common test score of 500 or better. The remainder were grandfathered in.1

Prior to the inauguration of the NTE, the faculty in the School District was 59 percent black, but after it was implemented, the percentage dropped to 52 percent by 1972. At the end of the 1970-71 school year twenty-five teachers were not offered new contracts and of those, twenty-one were black. Fifteen of these black teachers were terminated solely on the basis of their NTE scores.2 Twelve of them had from two to twelve years of teaching experience while three had only one year. However, all fifteen had been recommended for retention by their respective principals, several with above average or better overall ratings. In 1970-71, 127 in-service teachers were required to take the NTE test, twenty-one were black and 106 white. Seventy percent of the blacks failed, while only two percent of the whites did so. Only two white teachers were not re-employed as a result of an unsatisfactory score on the NTE. During this same year, the Board employed nineteen white and nine black teachers, none of whom had a college degree nor took the NTE test. In 1970-71, thirty-two white and seven black new teachers were employed who held 'collegiate certificates', a lower level certificate than the 'collegiate professional' certificate (awarded to teachers who had completed an education training program) held by the dismissed black teachers. The District Court noted that teaching outside a teacher's area of certification was an additional ground for dismissal; yet it appears that the Board employed thirty-six white and thirty-two black teachers for the 1970-71 school year to teach academic subjects outside of those for which they were certified. For the 1971-72 school year, the number of black teachers was further reduced by thirty-one, while the number of whites increased by twenty-seven; eighty-two new white teachers were employed but only fifteen new black ones.

The statistics on employment of new teachers was thirty eight percent black in 1970-71, while it was only fourteen percent in 1971-72. Significantly, we note that of the appellants whose employment was terminated for insufficient NTE scores, some had been teaching for over a decade and none had been teaching less than one year. For example, Celestine Whitehead had taught in Nansemond County from 1966-68 when she resigned to join her soldier husband stationed in Europe. Upon returning and during the 1970-71 school year she received the Teacher of the Year Award from the Chuckatuck Ruritan Club. She was terminated that year because of a deficient NTE score, although her principal rated her outstanding or above average in all evaluation categories and recommended the renewal of her contract. Another teacher, Josephine Gatling, was employed to teach physical education and driver education, subjects that were exempted from the NTE test under the announced policy guidelines of the Board. Nevertheless, she was required to take the test and upon her failure to make a score of 500 she was terminated at the end of the 1970-71 school year. On the other hand, other teachers in brick masonry, auto mechanics and electronics were exempted. Another teacher, Thelma Corprew, had taught three years, the last year in Nansemond County. She was terminated in 1971 because of her NTE score. She took the NTE test again, making 505, but was told the faculty had no vacancies; yet three white teachers were employed to teach in the same grade after she had been refused.

The trial court found that 'there was no racial motivation involved in any decision of the school board' but held that this 'lack of intentional discrimination' was not dispositive of the case although it acknowledged that two situations involving de facto discrimination in violation of individual rights were presented by the facts. The first was that the adoption of the NTE requirement resulted in the severance of more blacks than whites and secondly, that a 'larger number of blacks were demoted or refused re-employment based upon the subjective evaluation form and the recommendation of the principal.' The question, the Court posed, was whether the employment standards utilized by the Board were justifiable under the Equal Protection Clause. Citing Western Addition Community Organization v. Alioto, D.C., 330 F. Supp. 536 (1971), the Court reasoned that if there was 'a reasonably necessary connection between the qualities tested . . . and the actual requirements of the job to be performed,' (at 539) 'it would seem to make no difference whether the classification being attacked' was 'on the basis of race . . . The test need only reasonably measure those abilities which are essential to the job to be performed.' 351 F. Supp. at 203. The Court then applied this 'reasonable measure' criterion and decided that, although the NTE had no 'predictive validity', it did have 'content validity-- that is to say, the knowledge and abilities tested are those used in the job contemplated.' It then found that the evidence, 'including that from the plaintiffs' own experts, would seem to indicate that the test validly measures skills essential to the teaching profession.' Finally, the Court adopted, absent any clearly conflicting evidence to the contrary, the views of one of the expert witnesses, Dr. Deneen, who 'commented that many school districts abused teacher rights with an indiscriminate use of the NTE, having made no effort to justify its actions. Nansemond County on the other hand has what he considers to be faults but with good explanations for the use they have made of NTE.' Id. at 205. After thus approving the School Board's NTE policy, the Court discussed some of the individual teachers released, including those in which the principals had not recommended their retention and found justification for termination present in all cases save one who was terminated by the unification process of the School District. Her demotion was ordered by the Board without any comparison of her qualifications with others in the School District performing similar duties. She was given relief 'if the government has standing to seek relief in her behalf.3 For the reasons hereafter set out, we reverse the judgment and direct that the teachers terminated solely upon the basis of NTE scores be reinstated with full back pay; that the cases of the teachers terminated 'for cause' be re-examined at a hearing to be held by the District Court at the earliest practicable date to consider the validity of such dismissals and the appropriateness of reinstatement; that appropriate injunctive relief be issued and that damages, if appropriate, be awarded.

The issue as framed by the District Court is whether the Board's employment standards are 'justifiable . . . in light of the equal protection clause . . .'. It found that: 'there was no racial motivation involved in any decision of the school board, other than the avowed purpose of keeping the black to white ratio constant.' In our view this finding and the court's analysis overlook the rule of this Circuit in Chambers v. Hendersonville City Board of Education, 4 Cir., 364 F.2d 189 (1966), and which was recently approved and endorsed in Keyes v. School District No. 1,413 U.S. 189, 93 S. Ct. 2686, 37 L. Ed. 2d 548 (1973). There it was held that when the ranks of black teachers in a school district with a long history of discrimination have been decimated disproportionately, an inference of discrimination is raised which shifts to the school board the burden of justifying its conduct by clear and convincing evidence. As Mr. Justice Brennan said in Keyes, supra, 'In discharging that burden, it is not enough, of course, that the school authorities rely upon some allegedly logical, racially neutral explanations for their actions. Their burden is to adduce proof sufficient to support a finding that segregative intent was not among the factors that motivated their actions.' 413 U.S. at 210, 93 S.Ct. at 2698. The Board argues that no inference of discriminatory motive was raised 'because there were valid reasons for the use of the test.' We find the reasons articulated insupportable. The District Court did hold that for the NTE to pass muster, it was necessary that there be a 'reasonably necessary connection between the qualities tested . . . and the actual requirements of the job to be performed.' But we find this to be error. The Supreme Court in Griggs v. Duke Power Co., 401 U.S. 424, 91 S. Ct. 849, 28 L. Ed. 2d 158 (1971) held that the test must 'bear a demonstrable relationship to successful performance of the jobs for which it was used.' 401 U.S. at 431, 91 S.Ct. at 853. This standard is much more rigorous in its burden than the standard actually utilized by the District Court here. Using this less stringent standard the Court concluded that the NTE test had 'content validity' and could therefore be used in the teacher evaluation and selection process. We find no clear evidence to support that conclusion. The test itself was not introduced into evidence. Nor were there any comparative analyses or studies conclusively demonstrating that the knowledge necessary for the teaching positions bore any relation to the questions on the examinations.4 In this connection, the District Court itself found that 'the evidence clearly establishes that no study has shown any correlation between any score and an ultimately effective in-service teacher.' 351 F. Supp. at 203. If this be true, the NTE cut-off score of 500 is patently arbitrary and discriminatory. Superintendent Wood testified that he knew that black teachers were less likely to score as high as white teachers. Further, the reason for its adoption appears to be arbitrary based on the fact that Nansemond School District was getting the cast-off teachers of other school districts that required the 500 score cut off on the NTE. Mr. Wood testified that it was adopted because the Board 'found that we were getting many teachers coming in from Chesapeake, Portsmouth, Norfolk, Virginia Beach' and 'many of the teachers we were getting were unable to become certified in North Carolina'; and 'we found that all of these school divisions required a minimum of 500 on the National Teachers Examination . . . So that presented a problem . . . So taking all of this into account we felt that our teachers coming into Nansemond County should at least bring a minimal amount of general knowledge into theclassroom.' And the Board adopted the NTE 500 cut-off score to accomplish this. Moreover the Board compounded the problem of disparate racial impact by making the NTE 500 cut-off score the sole criterion for job selection, although the Educational Testing Service itself, the District Court found, 'could not condone blanket use of a minimum cut off score.' In addition, the Board arbitrarily and erratically administered the NTE, requiring some 'exempted' teachers to take the test and not requiring that others similarly situated do so.5 Finally, the District Court found, the Educational Testing Service 'has said that the test was least valid when applied to experienced teachers'. Yet the Board nevertheless required many teachers regardless of their experience (which exceeded ten years in some instances) to take the test and then refused them employment or re-employment if they failed to make a score of 500 or better. On top of all this, the Board employed twenty eight teachers in the school year 1970-71 and twenty four in the year 1971-72 without requiring either a college degree, which was the old requirement, or requiring them to take the NTE, which was the new one. Obviously, the District Court's finding that the action of the Board was not discriminatory is clearly erroneous and we so hold. As Judge Dyer said in Baker v. Columbus Municipal Separate School District, 462 F.2d 1112 (5th Cir. 1972) when the NTE 'produces a racial distortion it is subject to strict scrutiny . . . In order to withstand an equal protection attack it must be justified by an overriding purpose independent of its racial effects.' At 1114. We find no such overriding purpose here. While a school board's desire to improve the caliber of its faculty is a laudable one, the policies and procedures employed must be clearly and fairly related to this goal. The totality of the record, on the contrary, indicates a strong undercurrent of discrimination that has effectively decimated the ranks of black teachers whose credentials are equally, if not more, impressive than many of those hired by the Board.

We need not discuss at length whether the NTE has 'content validity'. The District Court noted in Footnote 8, 351 F. Supp. at 204, content validity alone may be acceptable where a well developed test, consisting of suitable samples of essential knowledge, skills or behaviors composing the jobs in question, are posed in the examination. 29 CFR 1607.5(a). However, we do not find the necessary prerequisites for such a finding by the District Court since no such samples, job analyses or validation studies were provided or conducted. Nor did any expert testify that the questions posed tested the existence of the knowledge essential to performance in the teaching positions. More critical is the District Court's express finding that the NTE had no 'predictive validity'. The burden of showing 'a manifest relationship' between the test and the job is upon the Board. Griggs v. Duke Power Co., supra, 401 U.S. at 432, 91 S. Ct. 849, 28 L. Ed. 2d 158. In that case the Chief Justice pointed out the inadequacy of 'broad and general testing devices.' Id., 401 U.S. at 433, 91 S. Ct. 849, 28 L. Ed. 2d 158. As concretely put by the Chief Justice, 'any tests used must measure the person for the job and not the person in the abstract.' Id., 401 U.S. at 436, 91 S. Ct. at 856. Here the tests measure the person in the abstract. While a copy of the National Teacher Examinations is not in the record, we have taken the liberty of reproducing in an Appendix some of the questions which appeared in the amicus curiae brief of National Education Association. If these questions are a fair example of the remainder of the examination, any connection between the examination and effective teaching is purely coincidental. The NTE, as used by the Board, does not purport to measure the teacher's actual knowledge of the subject matter assigned to be taught or his performance in the classroom, but places primary emphasis on general education and professional education.

In construing Title VII of the Civil Rights Act of 1964, the Supreme Court in Griggs, supra, stated unequivocally that: 'If an employment practice which operates to exclude Negroes cannot be shown to be related to job performance, the practice is prohibited.' 401 U.S. at 431, 91 S. Ct. at 853. Here the School Board conducted no 'meaningful study' of the relationship between the test and 'job performance ability.' Id. Since, as previously stated, the District Court expressly found that the NTE had no 'predictive validity' and since the statistics in the record confirm that Negroes have been excluded, we are led to the inescapable conclusion that the Board's use of the NTE was improper.

The blanket grouping of employees for testing in a single examination has been condemned by this Court in Moody v. Albemarle Paper Company, 474 F.2d 134, 139 (4th Cir. 1973). If the jobs are substantially different, it is reasonable to conclude that the relationship between jobs and the tests related thereto must vary as among various jobs. In the Nansemond schools some teachers might teach reading and writing to seven-year olds; others chemistry or foreign language to sixteen-year olds; still others may have classes in science, English, mathematics etc. Some of these subjects do not require much 'general knowledge' for one to be an expert in teaching them. A job analysis for one teaching position (and the appropriate test for it) would not necessarily be suitable for another. Of course, specific subject area testing is covered by another section of the examination called the Teaching Area Examination. Yet the Board did not require subject area testing. Had it done so, we surmise that experienced teachers might have scored more satisfactorily. In any event, it was unfair, we find, to give only the Weighted Common examination to experienced teachers who have long since finished their college careers. This requirement is contrary to the specific recommendation of the Education Testing Service. See Footnote 4, supra. General knowledge may be fleeting, and the Educational Testing Service designed the test, given by the Board, for new teachers who are finishing or who have just completed their college education. As the guidelines of the Educational Testing Service say: 'Test scores contribute little or nothing to the evaluation of an in-service teacher.' Hence the Board by requiring appellant teachers to take the NTE violated the direct recommendation of the Educational Testing Service itself.

Again, we must find fault with the NTE used by the Board for predictive purposes. The Board used a cut off test score of 500 as a prediction of the effective teaching capacity of its teachers. Indeed, Superintendent Wood testified that was why the Board adopted it; yet, the 'Prospectus' of the ETS states that the NTE is 'not intended as a measure of classroom teaching performance.'

We recognize that school boards should and do have wide discretion in the performance of the important duties assigned to them. Theirs is the most vital function in a free society-- the education of youth. The level of their accomplishment will determine whether a free society will long exist; for educated minds are the guardian genius of democracy. The well-springs of education must therefore be maintained pure and strong. At the same time the principle of equal employment opportunity is the law of the land (42 U.S.C. 2000e-2) and it must never be dishonored. To maintain the high purposes of education, we must encourage and protect those who devote their lives to this worthy endeavor. To this end school boards must make certain that teachers are not the victims of abitrary, discriminatory action, both in their entry to and continuance in the profession. It, therefore, is the affirmative duty of every school board to eradicate 'root and branch' every vestige of segregation and discrimination in their respective school systems. Green v. County School Board, 391 U.S. 430, 438, 88 S. Ct. 1689, 20 L. Ed. 2d 716 (1968). This may be a difficult task but it must be done. Such challenges are not new to Americans. If we falter in meeting them, we reflect our inadequacy for greatness.

What we hold today is not to say that the NTE cannot be utilized under any circumstances as one possible type of 'objective criteria' under the rule in Chambers, supra. However, it cannot be used as a tool of discrimination; it cannot be used without proper validation studies and job analyses under the rule in Moody, supra. Nor can it be administered capriciously in derogation of the guidelines promulgated by the ETS so as to have a racially disparate impact. It may well be that when properly and fairly applied in appropriate situations, the NTE could qualify as having the 'demonstrable relationship', as required by Griggs, supra, so essential to ensure equal employment opportunity.

We hold that the NTE, as applied here, was discriminatory and that the teachers terminated because of their failure to make a 500 score on the test must be reinstated with back pay and their damages, if any, settled. The facts surrounding the dismissal of certain teachers 'for cause' must be re-examined by the District Court with a view to making certain that their dismissal was not linked to discriminatory action with the burden of proof on the School Board; and if it was, then appropriate relief should be afforded them. Injunctive relief must be granted in such terms as will insure that further discrimination in the employment and retention of teachers in the School District will not recur. Finally, the District Court, in the exercise of its sound discretion, may grant such other and further relief as it deems necessary and appropriate. The judgment is reversed and the cause remanded for further proceedings in keeping with this opinion.

It is so ordered.

The professional education section of the test asks questions such as:6

'8. One curriculum pattern involves introducing significant concepts during the first year of school and continuing their development at increasing levels of complexity in the following years. This type of curriculum plan for concept development is called the (A) spiral approach (B) open-ended approach (C) concrete to abstract approach (D) contrasting dimensions approach (E) deductive approach

'9. Which of the following statements best describes the nature of mental growth in the child? (A) Mental development begins at birth and parallels physical growth in rate and length in years. (B) Mental growth begins at birth and ends at death. (C) Mental growth in the normal child is progressive but somewhat uneven. (D) The length and extent of mental growth cannot be altered after birth. (E) Mental growth necessarily parallels social and emotional development.'

The test of written English expression, for example, asks which, if any, of the following is in error:

'When it snows on/A the flat, open plains, it/B piles up into/C huge drifts that make/D roads impassable. No error/E and which of the following constitutes the best answer: 'The chemical formula was so complex, and no one was able to remember it. (A) so complex, and (B) very complex, therefore (C) so complex that (D) too complex that (E) of a degree of complexity so great that'

The remaining 50 percent of the examination calls for information such as the family living next door is not a 'social group' (Question 164); Ramsey Lewis is not a trumpet player (Question 188); St. Basil's Cathedral is an example of the Byzantine influence on architecture (Question 198); a Minotaur is half-man, half-bull (Question 208); the subject matter of an oratorio is typically religious (Question 215); Althea, Julia, Lucasta, and Corinna were ladies who inspired the poetic ardors of the Cavalier poets (Question 218); carbon dioxide is an excellent material for putting out fires because it is heavier than air and does not support combustion (Question 230); the incidence per capita of trichinosis is greater in the United States than in Asia because consumption of pork is low in Asia (Question 240); not all planets in the solar system have moons (Question 243); stars of the first magnitude are necessarily similar in brightness (Question 244); a circle circumscribed by a square, the area of which is 36, has as its own area (Question 262); for all integers x and nonzero integers y, the new symbol, when defined as, means that the following must be true. (Question 265).