Certain class representatives and dissident members of a class certified under Title VII and 42 U.S.C. Sec. 1981 ask that we overturn an approval by the trial court of a class settlement. Finding that the trial court did not abuse its discretion in approving the settlement and articulated its reasons for doing so with clarity and completeness, we affirm.

In March 1976, William Reed and others, on behalf of a class of black workers, filed suit against General Motors asserting claims under Title VII of the Civil Rights Act of 1964 and 42 U.S.C. Sec. 1981. A class certification hearing was finally held in January 1979. Before a ruling the case was transferred to Judge Barefoot Sanders, who after another hearing, certified on April 18, 1980, a class of:

All Negroes who have been employed by Defendant at the General Motors Arlington Assembly Plant on any date since September 17, 1970, who have been discriminated against because of their race by Defendant in any of the following terms and conditions of employment:

(1) Job Placement;

(2) Promotion;

(3) Transfers;

(4) Work Assignments; and

(5) Discipline.

The class shall not include former Negro employees who have not been employed by Defendant on any date since September 17, 1970.

Some fifteen months later and after extensive discovery all counsel requested approval and entry of a proposed consent decree in settlement of the case. The district court tentatively approved the proposed decree, directed notice to the class, and scheduled a class hearing. Notices of hearing and proposed settlement were mailed to all class members. Only 148 class notices of the 1,517 mailed were returned unclaimed. Presumably, 1,469 class members received actual notice. The proposed settlement immediately encountered opposition. By the time of the hearing, over 600 members of the class, and twenty-three of the twenty-seven named plaintiffs, had filed objections. At the hearing the objectors appeared through their own counsel and were given complete opportunity to express their opposition. Approximately three months later, on December 22, 1981, the district court approved the settlement in a written opinion. Reed v. General Motors, 560 F. Supp. 60 (N.D.Tex.1982).

Under the settlement, General Motors agreed to pay $200,000 to the class and promised not to discriminate against blacks. The $200,000 was to be distributed to class members under a point system that measured seniority. No injunctive relief was called for by the settlement. Contending that the relief was inadequate the dissident class members ask us to set aside the district court's approval of the settlement. We turn to the standard of review before returning to the specifics of the settlement.

Over the past twelve years, this court has, on at least five occasions, examined the requirements for approval of class action settlement and the measure to be employed on its appellate review. Parker v. Anderson, 667 F.2d 1204 (5th Cir.1982); In re Corrugated Container Antitrust Litigation, 643 F.2d 195 (5th Cir.1981); Pettway v. American Cast Iron Pipe Co., 576 F.2d 1157 (5th Cir.1978), cert. denied, 439 U.S. 1115, 99 S. Ct. 1020, 59 L. Ed. 2d 74 (1979); Cotton v. Hinton, 559 F.2d 1326 (5th Cir.1977); Young v. Katz, 447 F.2d 431 (5th Cir.1971). The teaching of these cases is that the district court's approval of a proposed settlement may not be overturned on appeal absent an abuse of discretion. See Young v. Katz, 447 F.2d at 432. The exercise of discretion is to be tested by inquiries that "ensure that the settlement is in the interest of the class, does not unfairly impinge on the rights and interests of dissenters, and does not merely mantle oppression." Pettway v. American Cast Iron Pipe Co., 576 F.2d at 1214. There are six focal facets: (1) the existence of fraud or collusion behind the settlement; (2) the complexity, expense, and likely duration of the litigation; (3) the stage of the proceedings and the amount of discovery completed; (4) the probability of plaintiffs' success on the merits; (5) the range of possible recovery; and (6) the opinions of the class counsel, class representatives, and absent class members. Parker v. Anderson, 667 F.2d at 1209. See also In re Corrugated Container Antitrust Litigation, 643 F.2d at 217.

The objectors here do not contend that the settlement was motivated by fraud or collision. Similarly, they do not dispute either the district court's finding that the settlement occurred after completion of discovery or the class counsel's testimony that the trial would be long and complex. Instead, the objectors argue that the product of the last three inquiries demonstrates that the trial court abused its discretion in approving the proposed settlement. We agree that these are the relevant inquiries but in making them conclude there was no abuse of discretion.

This inquiry contains an internal tension. A district court faced with a proposed settlement must compare its terms with the likely rewards the class would have received following a successful trial of the case. See Cotton v. Hinton, 559 F.2d at 1330. The court, however, must not try the case in the settlement hearings because "[t]he very purpose of the compromise is to avoid the delay and expense of such a trial." Young v. Katz, 447 F.2d at 433. Following this command, the district court here examined the pretrial record, conducted full hearings, and concluded that "if the case went to trial the class would have a good chance of proving its claim of discrimination in the areas of job placement and promotion." It also noted that the class would have little chance of showing discrimination in the areas of work assignment and discipline.

At the class certification hearing, plaintiffs supported the claim of discrimination in job placement, transfer, and promotion by showing an underrepresentation of blacks in salaried, supervisory positions. Specifically, plaintiffs showed that in 1970, blacks were 13.42% of the work-force but only .59% of the sixth level supervisors and that in 1976 they were 15.45% of the work-force but only 4.2% of the sixth level supervisors. These uncontroverted statistics are evidence that General Motors' placement, transfer, and promotion policies had a disproportionate adverse impact on blacks. See Griggs v. Duke Power Co., 401 U.S. 424, 91 S. Ct. 849, 28 L. Ed. 2d 158 (1971). If not explained they may have supported a finding of liability. We do not pause, however, to explore that possibility. We observe only that a finding of liability was by no means preordained, even if the figures were themselves unchallenged. In Pouncy v. Prudential Insurance Company of America, 668 F.2d 795 (5th Cir.1982), we suggested that a plaintiff must, as a part of his prima facie case, relate disparate impact to an employment practice. The fit of this decision in the jurisprudence of this circuit is less than obvious. Defining that fit is not our task here. Our reference to Pouncy is to illustrate only that this area of law is uncertain, continues to develop, and that its uncertainty was relevant to the settlement. For example, we are yet to absorb the technique of Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 101 S. Ct. 1089, 67 L. Ed. 2d 207 (1981) and General Telephone Company of the Southwest v. Falcon, --- U.S. ----, 102 S. Ct. 2364, 72 L. Ed. 2d 740 (1982). Specifically, their impact upon disparate impact cases is yet to be developed.

Finally, we note that plaintiffs provided little evidence of illegality in the areas of work assignment and discipline. In fact, the class counsel testified at the settlement hearing that he could find little support for these claims and the objectors have identified none, below or here. The district court thus did not abuse its discretion in its focus upon the areas of placement, transfer, and promotion.

In analyzing the fairness and adequacy of the $200,000 settlement figure, the district court multiplied the salary differential between a sixth level supervisor and an hourly worker by the deficiency in black supervisors for each year. After adding the yearly totals and making some minor adjustments, the district court concluded that the potential back pay recovery for discrimination in the areas of job placement, promotion, and transfer was in the $140,000 range. Its conclusion that "the settlement substantially remedies any class injury" is then unassailable when one adds the fact that there was no proof of pay differences attributable to job placement and transfer.

The objectors argue, however, that the $140,000 figure is too low because it does not reflect salary differentials between hourly workers and salaried, non-supervisory personnel. This argument is without factual support. Assuming that blacks were denied promotions to salaried, non-supervisory positions, the objectors failed to show any difference between the salary of non-supervisory personnel and the hourly wages to other employees under the collective bargaining agreement. In fact, the class counsel suggested that there is no difference. The objectors below and here have not shown to the contrary. The district court thus did not err in its assessment of the plaintiffs' range of possible recovery.

The more difficult question is whether the trial court abused its discretion in approving a settlement that does not provide for remedial or injunctive relief in the face of the court's finding that the class would "have a good chance ... in the areas of job placement and promotion." The settlement contained only General Motors' affirmative representation "not to discriminate against Blacks in regard to job placement, promotion, transfers, work assignments and discipline," a promise the objectors argue is inadequate.

The statistical presentation made to the trial court supports General Motors' claim of significant progress in reducing the disparity between similarly situated blacks and whites. In 1977 seven of twelve or 58% of employees promoted to sixth level supervisor were black despite a 18.51% work force percentage. The black percentage of sixth level supervisors grew from .59% in 1970 to 10.18% in 1979 with a growth in black percentage of work force of 13.42% to 19.06%.

By December 1969 the Arlington Plant had adopted an affirmative action plan. Pursuant to Executive Order 11246, that program was audited by the Office of Federal Contract Compliance Programs, Department of Labor. Approval letters were issued in the years 1972, 1973 and 1977.

In sum, there was record evidence that General Motors had made substantial progress and was continuing to do so pursuant to a written plan filed with and monitored by the federal government. Included in the changes are job postings and responses to the asserted practices of subjective personnel decision-making by white supervisors. See Rowe v. General Motors, 457 F.2d 348 (5th Cir.1972).

The quarrel of the objectors is ultimately that the settlement was improvident because it did not contain a general prohibition against discrimination. They do not suggest specific provisions that such a decree ought to have included. Putting to one side the practical value of such a generally phrased instruction as "don't violate the law," the question is whether the trial court abused its discretion in approving a settlement without its inclusion. In answering the question it is instructive to emphasize that this is not a case of a trial court refusing such remedial relief after an adjudication of liability. The absence of remedial relief was instead the product of the bargaining process. In that bargaining process both sides recognized that remedial relief had value. Even though it would have conferred little more than quick access to the federal court, its Damoclean force made it a desired objective for class counsel in the negotiation process. This observation regarding its worth, while intuitive, has marketplace support in the fact that class counsel sought it and General Motors resisted it. It proved to be an objective that was not achievable by agreement. As we have noted, the back pay award represented a substantial recovery with little discount for trial risk. The quid pro quo was rather plainly the absence of a general injunction with its threat of judicial superintendence of plant personnel decisions, a presence that would have been supplemental to that of the Office of Federal Contract Compliance.

There was no suggestion that any remedial relief, even if more than a general prohibition, would have been any more specific than the affirmative action plan in place. The negotiating concession of the class then was to give up an enforcing mechanism of an asserted violation of the court decree in return for an arguably less efficacious enforcement mechanism superintended by the Office of Federal Contract Compliance and a backpay pot sweetened by $60,000 over the provable award sum. Skilled and more than adequate lawyers for the class believed it to be a good bargain. The trial court did not abuse its discretion in approving it.

The objectors contend that the district court abused its discretion in approving the settlement over the objections of twenty-three of twenty-seven named plaintiffs and nearly forty percent of the 1,517 member class. They also argue that the nature of their objections required the district court to withhold its approval of the settlement. The district court carefully weighed both the number and nature of the objections in its evaluation of the settlement's fairness and explicitly recognized the controlling precedent.

"A settlement can be fair notwithstanding a large number of class members who oppose it." Cotton v. Hinton, 559 F.2d at 1331. At the same time, the number of objectors must be carefully considered. We undertook such an analysis in Pettway v. American Cast Iron Pipe Co., 576 F.2d at 1214-19. While we found there that vast class dissatisfaction with the settlement required the district court to withhold its approval, Pettway does not require a rejection of the settlement here. Pettway involved objections by all of the named plaintiffs and over seventy percent of the class. The dispute there centered around the sufficiency of the fund rather than its allocation and in this important sense was "an intrinsically 'class' decision in which majority sentiments should be given great weight." Id. at 1217. In contrast, many of the objections here involve the allocation of the settlement fund and were made by persons who were not economically disadvantaged and thus do not represent sentiments of the class as a whole. Under the allocation formula, approximately 900 class members would receive no money. Why this is so is apparent. No back pay was to be awarded after 1976 under the settlement formula because class counsel was unable to prove any denial of promotional opportunities by reason of race after that date. Because the statistics also proved the near impossibility of any promotion to sixth level supervisor without two years of seniority, persons without at least two years service would receive no money. If it is true as we find it to be that the allocation was a fair approximation of relative entitlement to back pay, the large number with no back pay entitlement suggests that the class may have been more broadly defined than appropriate; it is no suggestion that the settlement was not reasonable.

We pass from the assertion that the settlement was improvidently approved given the number of objectors to the quality of the lodged objections. In doing so we equally find no abuse by the trial court. The objectors first complain of the exclusion of persons without two years of service with General Motors and those who were hired after 1976--exclusions we earlier discussed and which were proper for the reasons stated.

The objectors' final complaint is that the settlement denies them the right to proceed against the United Auto Workers on their individual claims. It is not clear that the settlement has this effect. Regardless, the objectors did not mention it in either the written objections or at the fairness hearing. As such, this court will not consider the issue. See Mitchell v. M.D. Anderson Hospital, 679 F.2d 88, 91 (5th Cir.1982).

Backing away from the concerns of the immediate contestants, we note that class settlements of discrimination claims also implicate broad social concerns. Settlement of large numbers of discrimination claims serves the noble goals of Title VII and the Civil Rights Acts in a manner promoted by the traditional procedural insurers of industrial peace such as controlled bargaining and conciliation. Class litigation when not abused can aid the courts by its coagulation of numerous claims. It follows that it is relevant to ask whether a settlement results in a decision binding on the class or whether it acts as a dispersing agent. For example, the ability of the Pettway class members to opt out of the settlement and pursue individual claims counseled against approval of a disputed settlement. In contrast, class members here who did not earlier opt out of the class are bound by the settlement. In this sense the settlement reduces the burdens placed on the judicial system. Admittedly too broad to provide analytical heat, these heuristic expressions nonetheless shape judicial attitude and are properly the backdrop to the Parker inquiries.

In reviewing proposed class settlements, a trial judge is dependent upon a match of adversary talent because he cannot obtain the ultimate answers without trying the case. Indeed, that uncertainty is a catalyst of settlement. Because the trial judge must predict, the value of the assessment of able counsel negotiating at arm's length cannot be gainsaid. Lawyers know their strengths and they know where the bones are buried. The analytical construct of Parker built upon the cases that preceded it recognizes the crystal ball dimension of the trial judge's task and channels his predictive inquiries in a way believed best for accuracy. It is doubtlessly true, as Judge Tjoflat observed in Corrugated Container, that the settlement itself provides insight into adequacy of representation. Yet, this observation is only another way of asking the same question for the reason that adequacy of representation and adequacy of settlement are different sides of the same question. We reaffirm the Parker inquiries, emphasizing that the linchpin of an adequate settlement is adequacy of representation. That it is the trial judge who can best know how well the class was represented informs the discretion given to him in reviewing proffered settlements under Rule 23. Finding no abuse we here AFFIRM.