Eleanor Duke appeals from a judgment in a bench trial in favor of her employer, The University of Texas at El Paso. Duke, a biology professor at UTEP, brought suit under Title VII, 42 U.S.C. Secs. 2000e et seq., and 42 U.S.C. Sec. 1983, alleging sex discrimination in pay, promotion and teaching opportunities. On appeal, she argues that the district court prejudiced her effort to obtain class certification and to prove discrimination in her individual suit by unduly restricting discovery. We review the discovery rulings by an abuse of discretion standard and reverse.

* In July 1973, Duke filed a charge with the Equal Employment Opportunity Commission, alleging that UTEP was guilty of university-wide discrimination against female faculty members in pay, promotion and work assignments. An EEOC investigation found reasonable cause to believe that UTEP discriminated on a university wide basis in pay and promotion procedures. The EEOC then sued UTEP, alleging violations of the Equal Pay Act. 29 U.S.C. Sec. 206(d). On September 24, 1979, the EEOC notified Duke of her right to sue under Title VII.

On December 20, 1979, Duke sued UTEP, individually and as class representative of all female faculty members, alleging that UTEP, contrary to the constitution and Title VII, discriminated against her and other female faculty members in pay, promotion and teaching opportunities. On January 30, 1980, the district court issued its standard order that discovery be completed by all parties within four months, but within thirty-five days granted UTEP's motion to dismiss Duke's complaint for failure to exhaust her administrative remedies. Upon Duke's motion to reconsider, the district court held a hearing on March 27, 1980 and withdrew its order dismissing the complaint. It then held a second hearing on the motion to dismiss on August 11, 1980, at which it took evidence. On September 8, 1980, the district court granted summary judgment to UTEP, finding that Duke had failed to exhaust her administrative remedies. We reversed that judgment, 663 F.2d 522, and the mandate was filed with the District Court on February 9, 1982.

Seven days after remand, UTEP moved to consolidate Duke's case with a similar suit filed by Julie Hansen, a math professor at UTEP, arguing that the suits charging sex discrimination in the two departments presented common questions of law and fact. They were consolidated on February 22, 1982. On March 5, 1982, Duke sought further consolidation with the still-pending EEOC case. UTEP opposed this consolidation and Duke's motion was never ruled on, although the district judge presiding over the EEOC case denied a similar motion.

On February 23rd, approximately two weeks after this court's remand, the district judge set the case for trial on March 29, 1982. Duke moved to delay trial until she had a reasonable time to complete discovery, pointing out that, among other difficulties, no hearing on her requested class certification had been held. The district court removed the case from the trial setting.

With the case removed from the trial setting, discovery started. In response to Duke's April 1982 interrogatories inquiring about all departments that contained male and female faculty members, UTEP provided information concerning salary promotion and work opportunities for the biology and math departments but, on April 26, 1982, moved for a protective order confining discovery to those two departments. In seeking to confine the discovery UTEP argued that, because relevant personnel decisions were made at the department level, data about other departments would not lead to admissible evidence.

Four days later, on April 30, 1982, the district court sua sponte scheduled a class certification hearing for May 19, 1982. On May 13, 1982, Duke formally moved for class certification and undertook to brief the merits of UTEP's motion for protective order and prepare for the certification hearing without the requested discovery. At the May 19th hearing, the district judge refused to decide the questions concerning the scope of discovery but indicated that he would later decide those questions on the submitted briefs.

On June 14, 1982, the district court denyed Duke's motion for class certification. He didn't resolve the discovery dispute until October 18, 1982, the day before trial on Duke's individual claim, when he sustained UTEP's effort to confine discovery.

It is apparent that Duke was not given an adequate opportunity to conduct meaningful discovery before the class certification hearing. Despite UTEP's successful claim that common questions of law and fact existed between Duke, the biology professor, and Hansen, the math professor, the trial court refused to grant Duke access to promotion and pay records of other departments of the University and then refused to certify any class of female professors, finding insufficient common questions of law and fact.1 Similarly, in Duke's individual trial the court found that Duke failed to show that the reasons given for her disparate treatment were pretextual, see Texas Dept. of Community Affairs v. Burdine, 450 U.S. 248, 252-53, 101 S. Ct. 1089, 1093, 67 L. Ed. 2d 207 (1981); McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802-804, 93 S. Ct. 1817, 1824-25, 36 L. Ed. 2d 668 (1973). Indisputably, Duke's ability to show that UTEP's explanations were pretextual was central to her discrimination claim. That effort was prejudiced by the cropping of her opportunity to discover the practices of the other university departments.

It is significant that the record reflects that the discovery request was based on more than a fanciful hope of counsel. A UTEP report had found in an internal study that eleven of the then 69 female faculty members were paid less than their male counterparts because they were women; that self-study was not confined to the math and biology departments. While UTEP successfully argued to the district court that relevant decisions were made at a departmental level, Duke was not allowed the opportunity to place that asserted "fact" in context. We cannot say that this effort was here so improbable of success that all discovery from other than the biology and math departments ought to have been denied. The EEOC found, for example, that

Each department evaluates its own faculty and is free to develop its own methods and procedures for evaluation for tenure, promotion and salary increases. However, we found, from the record, that the method used in the Biological Science Department is fairly representative of the method used in Respondent University as a whole.

UTEP now urges that Duke's counsel ought to have capitalized on the discovery conducted in the EEOC suit. Counsel for Duke did attempt to participate informally in the discovery occurring in the EEOC case, but UTEP refused. Then at the October 18th hearing, when the motion for protective order was being decided, UTEP's counsel argued that any disadvantage incurred by Duke by not confining discovery to the math and biology departments was minimized by her ability to use the EEOC's "expert" who had attempted several regressions of data obtained in the EEOC case. Yet, at this same hearing UTEP's attorney also urged that the expert's testimony not be admitted because the data base on which he relied was hopelessly flawed. At trial UTEP pointed out that the expert's analysis contained 463 errors.

In sum, there is no question but that Duke's April 1982 discovery requests, which were pending before the district court in May 1982, were relevant to both the class action certification and Duke's individual claim. Under the circumstances, failure to allow the requested discovery was an abuse of discretion. Though we are reluctant to find abuse when the trial court's discretion is as wide as it is in the superintendence of discovery, we have done so where the facts mandated such a finding.

In Burns v. Thiokol Chemical Corp., 483 F.2d 300 (5th Cir.1973), we reversed a judgment in which Burns was found not to have been the victim of racial discrimination and remanded for a new trial because Burns had been denied discovery of data of employment practices throughout the plant at which he worked. We stated:

Because discovery matters are committed almost exclusively to the sound discretion of the trial Judge, appellate rulings delineating the bounds of discovery under the Rules are rare. But the Judge's discovery rulings, like his other procedural determinations, are not entirely sacrosanct. If he fails to adhere to the liberal spirit of the Rules, we must reverse. See Wallin v. Fuller and Nationwide Mutual Insurance Co., 5 Cir.1973, 476 F.2d 1204 [1973]. And this is especially true in Title VII cases where courts have refused to allow procedural technicalities to impede the full vindication of guaranteed rights. Sanchez v. Standard Brands, Inc., 5 Cir.1970, 431 F.2d 455, 461.

Burns v. Thiokol Chemical Corp., 483 F.2d at 304-05. See also Trevino v. Celanese Corp., 701 F.2d 397, 406 (5th Cir.1983).

Finally, UTEP misses the mark in suggesting that any prejudice to Duke was harmless because Duke later lost on trial of her individual claims and cannot complain on behalf of a class of which she was not a member. Had Duke's class claims prevailed, she would have faced a distinctively less onerous burden at the trial of her individual claim.

We are sympathetic with the trial court's efforts to contain discovery and manage a heavy docket. A trial judge must be able to direct discovery and act promptly to prevent abuse without being second-guessed by appellate judges remote from the scene. Here, however, we are persuaded that the curbs on discovery were not sufficiently tailored to the case and undue prejudice resulted. We reverse and remand to allow the parties to complete discovery, for a new hearing on requested class certification, for a trial of the claims of any certified class and for a new trial of Duke's individual claim.

REVERSED and REMANDED.