JAMES DICKSON PHILLIPS, Circuit Judge:
James B. Wright, a black civil service employee of the National Archives and Records Service appeals from a judgment denying any relief in his action brought under Title VII of the Civil Rights Act of 1964, 42 U.S.C. §§ 2000e-5(f)(3) and 2000e-16(c) alleging discrimination on the basis of race in certain personnel actions taken in relation to his federal employment.1 Accepted with two other black and one white fellow [706]*706employees into a training program designed to qualify the trainees for promotion to higher civil service grades, Wright was evaluated unqualified for a second promotion at its end. When he declined to accept a proffered special extension of training to see if he might later qualify for promotion, he was removed from the program. Alleging discriminatory personnel actions taken during the course of the training program and in the failure to promote him at its conclusion, he sought, with other relief, promotion and back pay at the higher grade. Following a bench trial, the district court concluded that no violation of Title VII had been established.2 We agree and affirm.3
I
The Archives Specialists Training Program was instituted in 1968 by the National Archives and Records Service at the Washington National Records Center to develop personnel to fill the positions of supervisory archives technician and archives specialist. At its outset, applicants were required to be graduates of a four-year college and to have passed the Federal Service Entrance Examination. The program covered a two-year period and applicants were required to enter at the grade level GS-5. Those who successfully completed the first year would be promoted to GS-7; those who successfully completed the second year to GS-9. Any applicant who had already attained a grade level higher than GS-5 had to drop back to GS-5 to enter the program. Successful graduates of the program would be assigned to positions as supervisory archives technicians or as archives specialists.
In 1969 an investigation into possible racial discrimination at the Washington Records Center was conducted. A report documenting the existence of discrimination resulted. Following this, the requirements that applicants to the training program be college graduates and have passed the Federal Service Entrance Examination were dropped. Thereafter, black employees of the Archives and Records Service were actively recruited for the training program as part of the Agency’s Equal Employment Opportunity Affirmative Action Program. Among those recruited was plaintiff James B. Wright.
On April 19, 1970, Wright signed a training agreement and entered the program. He was not a college graduate. Since 1957, he had been employed by the General Services Administration and the National Archives and Records Service. Classified as a GS-6, he had to accept a grade reduction to GS-5. He was the first black to enter the program. In May 1970, three other members of the training class were accepted: Ernest Nolan, a black college graduate; Robert Sistare, a black without a college degree; and Orace Whitelock, a white with a college degree.
During the progress of this “1970 class” of four through the training program there occurred a series of incidents which, though not all central to Wright’s specific claims of discrimination, bear significantly upon resolution of those claims. The first involved work space assigned to the trainees. White-lock, the sole white trainee, was the first of the class to begin the program. When he arrived, the area where the trainees were to work was under construction, so he sat in the same cubicle with other trainees from earlier classes, as directed by a superior. When Wright arrived, he asserts that he either had to erect his own cubicle or was assigned to one that had just been constructed by someone else. The other black trainees were seated with Wright. After the trainee or trainees (the number seems to have varied) with whom Whitelock sat [707]*707had graduated, Whitelock had his cubicle to himself. After the blacks complained about the apparently segregated working space conditions, Whitelock’s desk was moved into the same vicinity as theirs. This was some two or three months after the program began.
The second incident involved the procedures by which the trainees were rotated through the various branches of the Service. During the course of the program, each of the trainees was supposed to rotate through each of the branches twice. Wright complains that whereas Whitelock was put into the rotation immediately, he and Sistare were not. Be that as it may, the design was generally followed. Wright certainly made the full rotation twice. The type of work assigned to the trainees in each branch was apparently determined by whoever was supervising them at the time and probably depended on what was available to be done. The trainees did not all follow the same rotation schedule. Wright contends that Whitelock received more challenging assignments which better prepared him for a managerial position. The Manpower Officer for the program, Wendell Evans, testified that Whitelock may have been given tougher assignments because he was doing better work and progressing faster than the other trainees.
The next set of incidents involved the quality of supervision and evaluation of the trainees provided by their several supervisors, particularly defendant Jean Fraley. Wright adduced evidence that in a variety of ways the supervisors gave less instruction to the black trainees than to Whitelock and that they were more reluctant to take the time to answer questions from the blacks than from Whitelock. Whitelock essentially acknowledged that he received closer supervision, but asserted that this was because he took the initiative to seek this out.
Fraley was the supervisor under whom Wright worked when he began rotating through the various branches of the Records Center. The black trainees perceived her to be racially prejudiced and some non-blacks also did. She possessed what was described on trial as an “abrasive” personality and some who testified attributed her behavior to that. The district court found, and this is not controverted on appeal, that Fraley gave closer instruction and supervision to Whitelock while he worked under her than to the black trainees. This, however, was found to be the product of a mistaken belief on Fraley’s part that White-lock was to be permanently assigned to her branch. After the black trainees complained, Fraley was removed from her position.
At the end of his first year, Wright was promoted to GS-7 as were the other trainees. Until well into his second year in the training program his evaluations were “good” or “excellent” except for one evaluation of “poor.” When he attempted to determine who gave it, he was at first unable to obtain an answer, but eventually learned it was Fraley. He then sought and was successful in having Fraley’s superior, Walter Barbash, raise it to a “fair.” During the latter period of his second year in the program, Wright’s evaluations began to fall off. The evidence as to cause is in sharp conflict. Wright contends that it resulted from a stiffening of the evaluation criteria at that time. Another explanation, of which the record contains only a suggestion, is that during the earlier stages of the program someone was assisting Wright in preparing his reports and that when that stopped his deficiencies appeared. No specific findings of fact as to this matter were made.
The ultimate focus of Wright’s claim of discrimination relates to unequal special training opportunities provided the three black trainees as opposed to those provided Whitelock. In addition to rotation through the various branches of the Records Center, the trainees were provided with a number of courses to provide special training. To some extent these seem to have been tailored to the perceived needs of each trainee. Wright, who had writing problems, asked for and was assigned a composition course in the graduate school of the Department of [708]*708Agriculture. Whitelock attended courses in “Work Force Estimating” and “Work Simplification Techniques.” From this, it is contended that Whitelock attended courses that were better adapted to training managerial skills.
The main contention of discriminatory unequal training however pertains to a so-called Special Projects Staff created in May 1971 to serve the Washington National Records Center. Twice recommended in the late 1960’s, its principal function was envisioned to be that of dealing with management problems arising throughout the Records Center. Originally the plans had called for a staff of nine, but due to economic concerns, Joseph Bradt, then Manager of the Center and a defendant in this action, asked only for a chief and six staff members. Only the chief and three staff positions were created; and the defendant Roy Wilson was assigned as Chief of the Staff. Bradt, Wilson and Walter Stender, Assistant Archivist for Federal Records Centers, were charged with selection of the original staff. The three people selected were a recent graduate of another special training program, a recent graduate of the program in which Wright was a participant, and Orace Whitelock. All three were white.
The stated basis for Whitelock’s selection was that the Special Projects Staff, whose implementation had been delayed several years, had a number of projects of a pressing nature which would require immediate attention. Due to the small staff, supervision would have to be minimal. Whitelock had performed better than the other trainees, was a superior writer and could work with less supervision. In selecting White-lock for the position, Bradt did not consider rotating all of the trainees through the Special Projects Staff. However, as events developed, each of the black trainees, including Wright, rotated through the Staff as trainees for one month each and in late 1971 or early 1972, they were regularly assigned to the Staff under circumstances that will later be described.
The length of time that Whitelock served on the Special Projects Staff when the black trainees did not is the subject of dispute. Wright contends it was six or seven months; defendants claim it was closer to three. Whitelock, according to his own quarterly reports, was appointed to the staff on May 3, 1971 and he immediately assumed his duties there. The exact date upon which the black trainees were formally assigned to the Staff is not clear, but Wright began his tasks there on December 23, 1971 and Nolan and Sistare seem to have taken up their duties with Special Projects at about the same time. In addition it is conceded that Wright spent one month on rotational assignment to Special Projects between May and December, but the record does not indicate when. It is of course obvious that, whatever the exact time, Whitelock had the experience provided by this assignment for an appreciably longer portion of the two-year training program than did Wright and the other trainees.
In the fall of 1971, the defendant Forest Williams replaced Bradt as the Manager of the Washington Records Center. He had a number of misgivings about the way in which the training program was being conducted and about the use of the program as an affirmative action project. He felt that the trainees were spending too much time on operations-oriented projects as opposed to tasks that were more managerial-oriented. When the trainees were not on rotation they tended to slip back into performing the duties they had performed before being assigned to the training program. He believed that the evaluations of the trainees had been overly generous and that the second year, and particularly the last six months, of the program should be more demanding than the preceding portions. He also expressed doubts about the viability of any program where everyone passed.
A related concern of Williams’ was that the training program was a poor vehicle for affirmative action. He testified that he had had qualms about it before becoming Center Manager, but that his belief had not crystalized until he gained some first-hand [709]*709knowledge. By selecting underskilled applicants, without diminishing the standards of performance demanded by the program, he believed the Agency might have dis-served the blacks selected.
Williams undertook a number of steps to improve the training program after he became Manager. He insisted that evaluations be more rigorous and that the trainees be assigned tasks that better prepared them for managerial positions. In late 1971 or early 1972 he assigned the black trainees to the Special Projects Staff and stated before trial that there were no money problems that precluded appointing anyone to that program. At trial, when the conflict between that statement and Bradt’s explanation for why only Whitelock had been appointed originally was brought to his attention, Williams clarified his position by stating that the money problems had been relieved by certain economic moves that had been taken in the summer and fall of 1971. When the blacks were assigned to Special Projects, a black supervisor, Edith Thomas, was also assigned there. The amount of supervision provided the black trainees was allegedly still minimal, but Nolan testified that it was more than they had been receiving previously. It is undisputed, indeed stipulated, that Whitelock’s training while he alone was on Special Projects Staff was superior to that being received at the same time by the black trainees, including Wright.
On March 1, 1972, Nolan dropped out of the program. The exact reason for this is not clear. He had discipline problems and was not performing well, and this is the presumable explanation. This left Wright, Sistare, and Whitelock.
On March 22, 1972, an evaluation panel met to consider whether Wright should be passed through the program. The five members were Forest Williams, Roy Wilson, Edith Thomas, Wendell Evans (the Manpower representative for the program), and Patricia Moore (the career development officer for the program). Three were blacks: Thomas, Evans and Moore; two were under the supervision of panelist Williams; Wilson and Thomas; all but Moore had supervised Wright during his training. During that meeting Wright was asked what he felt he needed to work on to become a successful manager. He replied “nothing.” The panel disbanded without making a recommendation.
The panel reconvened on March 27. Before that meeting Williams, Evans and Walter Stender, Assistant Archivist for Federal Records Center, had met privately and decided that Wright should be offered a ninety-day extension of the training program. The panel accepted that proposal unanimously, being of the opinion that Wright had failed to meet the minimum standards of accomplishment required for graduation. The offer was made to Wright.
A dispute exists with respect to what was said when Williams offered the extension to Wright. Wright claims Williams said the odds were ninety out of a hundred that he would not be promoted after that time. Williams attributes that statement to Wright and claims he replied that the odds were fifty-fifty. It is uncontroverted that Williams took the occasion to reiterate his belief that use of the training program as an affirmative action vehicle might be working a disservice to the minority participants. Wright declined to accept the extension. Apparently the offer was then made a second time and refused once more.
Williams then dropped Wright from the training program. At the objection of Moore, the panel was reconvened and voted three-to-two in favor of Williams’ unilateral action. The three votes were cast by Williams, Wilson and Thomas. Evans felt the panel should have made the initial decision whether to drop Wright, but has stated that he would have voted to drop him. Moore also believed the panel should have made the decision, but she would have voted to promote Wright.
Other than Moore, only Walter Barbash among the supervisors who testified felt that Wright was qualified for promotion to GS-9. Even Barbash testified that Wright’s competency decreased as the prob[710]*710lem he was working on called for more managerial skills. He further testified that the aim of the training program was not to produce graduates who would dead-end at GS-9. Rather the goal was to produce managers who could eventually function at levels from GS-12 to GS-14.
In mid-April, Sistare and Whitelock were promoted to GS-9 and permanently assigned to Special Projects. Sistare received special help from Wilson in completing his final project. The training program was then disbanded. Wright continued his employment in the Washington Records Center, and instituted these legal proceedings.
II
A.
It is necessary at the outset to identify the particular actions claimed by Wright to be discriminatory and to clarify the appropriate mode of legal analysis to be applied in assessing the claims. During the course of the litigation there may have been some confusion and shifting of positions as to each, a circumstance quite understandable in view of the evolving state of controlling doctrine during this protracted litigation. See Nashville Gas Co. v. Satty, 434 U.S. 136, 148, 98 S.Ct. 347, 54 L.Ed.2d 356 (1977) (Powell, J., concurring).
In his complaint Wright identified both the ultimate failure to recommend him for completion of the program with consequent promotion and various aspects and incidents of the training program itself as the discriminatory actions upon which his claims were based.4 The district court addressed both of these and found no discrimination either in the conduct of the training program or in the terminal denial of promotion. On this appeal Wright has focused essentially on allegedly discriminatory aspects of the training program, apparently conceding that if the failure to certify his completion of the program with consequent promotion were viewed in isolation he could not sustain a claim of racial discrimination in that specific action. Be that as it may, we have considered it appropriate to review the district court’s treatment of both sets of actions.
There has also been some confusion as to whether the claims are properly to be analyzed under the “disparate treatment” test of McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), or under the “disparate impact” test of Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971), or under both. The district court employed primarily a disparate treatment analysis of the proof, but its oral opinion can fairly be read as having also considered the proof, albeit summarily, under a dispa[711]*711rate impact analysis, finding the claims not established under either. On appeal the parties have eventually joined issue under both modes of analysis, and it is on this basis that we review the district court judgment. If any doubt has existed about the appropriate way to assess Title VII claims that seek to invoke both “disparate treatment” and “disparate impact” theories as alternative grounds for relief, Furnco Construction Corp. v. Waters, 438 U.S. 567, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978), decided during the pendency of this appeal, has made it clear through the mode of analysis employed in that case that these are rightly treated simply as alternative theories upon7 which a right to relief under Title VII may be established in a given case.5 As such, in keeping with general rules of civil procedure, they may properly be pleaded, Fed.R. Civ.P. 8(e)(2), and if sufficiently supported by evidence, Fed.R.Civ.P. 41(b), analyzed by the trier of fact as alternative grounds of relief. Obviously, as alternative theories and not separate claims they may not be applied to establish multiple violations on the same facts, but short of this no procedural election between them is required of a claimant at either pre-trial, trial, or appellate stages.6
On this basis, we analyze the claims, as did the district court, under both theories, looking essentially to the contrasting definitions of the two given in International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977). There, disparate treatment is said to define a situation where “[t]he employer simply treats some people less favorably than others because of their race, color, religion, sex, or national origin.” Id. at 335 n.15, 97 S.Ct. at 1854. On the other hand a disparate impact situation is defined as one that “involve[s] employment practices that are facially neutral in their treatment of different groups but that in fact fall more harshly on one group than another and cannot be justified by business necessity.” Id.
We take the disparate impact analysis first.7
B.
As indicated, the district court considered and rejected in summary fashion the possibility that Wright had established a right to relief under a disparate impact theory.8 We agree. Central to proof of a prima facie case under this theory is proof of an employment policy or practice which, though facially neutral or even benign in actual purpose, nevertheless imposes a substantially disproportionate burden upon a claimant’s protected group as compared to a favored group within the total set of persons to whom it is applied.9 Nashville Gas [712]*712Co. v. Satty, 434 U.S. 136, 98 S.Ct. 347, 54 L.Ed.2d 356 (1977); Griggs v. Duke Power Co., 401 U.S. at 430-32, 91 S.Ct. 849. Wright would have us find such a policy or practice and such a disparate impact here in various “actions which were supposed to be facially neutral but which resulted in the black trainees’ [sic] receiving inferior and unequal training and thus had a ‘disparate impact’ on the black trainees.” Reply Brief at 13. Only by a tortured logic that would completely wrench “policy or practice” and “disproportionate impact” from their common and acquired legal meanings in the employment discrimination context could any policy or practice causing disparate impact be found here. The set of persons affected by the personnel actions in question was four in number; of the set of four, three were members of claimant’s protected minority group; and of these three, only one is alleged finally to have suffered any legally cognizable harm from the actions charged to be discriminatory.10
Within this setting Wright’s legal claim, advanced solely on his behalf and not that of any other black trainee, thus comes to the proposition that the training program as administered did not equip him, a black, but did equip Whitelock, a white, for the promotion in grade that was the goal of both and the purpose of the program. If this is to be found a violation of Title VII under disparate impact doctrine, the “policy or practice” must be found in the several discrete decisions that ordered the sum total of the training received by these two, and a “disparate impact” in the different results finally experienced by just these two. This simply cannot be done except by warping the doctrine beyond recognition.
The policy or practice contemplated by disparate impact doctrine consists of “more than the mere occurrence of isolated or ‘accidental’ or sporadic discriminatory acts,” having reference instead to an employer’s “standard operating procedure— the regular rather than the unusual practice.” International Brotherhood of Teamsters v. United States, 431 U.S. at 336, 97 S.Ct. at 1855 (defining “pattern or practice” in Government action under § 707(a) of Title VII). Though the dividing line between a series of “isolated acts” and “standard operating procedure” will obviously sometimes be hard to draw, we think that on any fair and common sense appraisal the actions complained of here must be characterized as the former and not the latter.
While we quite agree that an individual may, in appropriate circumstances establish without elaborate statistical proof a disparate impact prima facie case, see Mitchell v. Board of Trustees, 599 F.2d 582, 585 n.7 (4th Cir. 1979), we cannot agree that the doctrine can fairly be applied to a situation such as that one here where the total set of affected persons is so small, the protected group is numerically dominant in the total set, and the adverse “impact” is finally experienced by only one of the numerically dominant protected group.
Disparate impact doctrine as first laid down in Griggs was clearly de[713]*713signed to ensure more perfect realization of the beneficent purposes of Title VII by making plain that discriminatory consequences as well as discriminatory intent fell under the ban of this remedial legislation, and by providing a relatively easy burden of prima facie proof of discriminatory consequences to overcome difficulties that might normally obtain in proving “discrimination in employment” vel non. As even a cursory survey of employment discrimination cases since Griggs will easily show, it has been a mighty instrument in correcting the more subtle and sometimes unconscious forms of discrimination that have pervaded employment as tragic legacies from our history. For this reason, it should not be applied by courts with a rigidity or inflexibility that effectively undercuts its intended function of fairly implementing Congressional purpose to root out these more subtle forms. By the same token, its integrity as a rational judicial construct to accomplish those purposes should not be compromised by undisciplined extensions that finally rob it of elemental predictability as an ordering principle for counseling, for voluntary corrective action by employers, and ultimately for litigation. In the long haul, such a warping of principle will not serve the Congressional purposes. To find a policy or practice giving rise to disparate impact on the facts of this case would be just such an undisciplined extension,11 and we decline to make it. We conclude instead that the evidence fails to establish a prima facie case of violation under the disparate impact theory,12 and turn now to the question whether violation may have been established under a disparate treatment analysis.
c.
The framework for analysis of Wright’s claim under a disparate treatment theory is that originally laid down in McDonnell Douglas, and subsequently refined and clarified in the course of applications in Furnco Construction Corp. v. Waters, 438 U.S. 567, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978), and Board of Trustees v. Sweeney, 439 U.S. 24, 99 S.Ct. 295, 58 L.Ed.2d 216 (1978) (per curiam), as adapted to the particular facts of the instant case. As the Teamsters Court pointed out in distinguishing this theory from the disparate impact theory, it finds violation where an “employer simply treats some people less favorably than others because of race”; which means that “[p]roof of discriminatory motive is critical, although it can in some situations be inferred from the mere fact of differences in treatment.” 431 U.S. at 335 n. 15, 97 S.Ct. at 1854. These cases establish a practical scheme of proof for addressing the critical issue of intentional racial or other impermissible discrimination in the treatment accorded a claimant. In its first stage, a claimant has “the initial burden of showing actions taken by the employer from which one can infer, if such actions remain unexplained, that it is more likely than not that such actions were ‘based on a discriminatory criterion illegal under the Act.’ ” Furnco Construction Corp. v. Waters, 438 U.S. at 576,. 98 S.Ct. at 2949. If the claimant meets * MESSAGE(S) *MORE SECTIONS FOLLOWthis production burden, a legal inference of discriminatory motive for the action is created that suffices to carry the claimant’s initial burden of production. This shifts to the employer the burden of production, of “going [714]*714forward”13 with evidence of “some legitimate, nondiscriminatory reason for the [action].” Id. at 578, 98 S.Ct. at 2950. This, if successfully done, will “dispel” the original adverse inference and entitle the employer to judgment unless in the final stage the claimant can carry the burden of proving “that the proffered justification is merely a pretext for discrimination.” Id. In this scheme, employing the familiar device of the presumption to establish a claimant’s prima facie case,14 the burden of persuasion of violation, including that on the critical issue of discriminatory motive, remains throughout upon the claimant. We conclude that the district court faithfully and accurately followed this proof scheme in analyzing the evidence and finding no violation.
As indicated in Part II.A. supra, Wright’s claim of violation as originally pleaded identified both the terminal failure to promote and the antecedent failures to provide equal training opportunities as the discriminatory actions upon which his claim was based. The unequal training incidents relied upon consisted essentially of White-lock’s being given better supervision and aid by all the supervisors, particularly Fraley; Whitelock’s being given more challenging assignments and better special training opportunities; and, most critically, Whitelock’s being given substantially more training experience with the Special Projects Staff than were Wright and the other black trainees. On the evidence adduced, analyzed within the proof scheme above outlined, we agree with the district court: (1) that a prima facie case was not established in respect of the terminal failure to promote, viewed as a discrete act of alleged discrimination; (2) that if a prima facie case was established in respect of the training incidents, viewed either separately or in total effect, it was effectively dispelled as to each by the “articulation of legitimate nondiscriminatory reasons” which in turn were not overcome by proof that they were merely “pretexts” for racial discrimination.
The McDonnell Douglas formula for proof of a prima facie case, as adapted to failure to promote, would require proof (1) that Wright belonged to a racial minority; (2) that he sought and was qualified for promotion to a position to which promotions were being offered to persons of his qualifications by his employer; (3) that despite his qualifications, he was rejected; and (4) that, after his rejection, the position remained available by promotion to others of his qualifications. The evidence clearly supported a finding that Wright was not qualified for the terminal promotion denied him. All five of the evaluators, three of whom were black, agreed in assessing Wright’s work that he was not qualified. The district court was entitled to accept this judgment as the basis of its own determination of nonqualification in the absence of any evidence drawing its essential integrity in question.
[715]*715As indicated earlier in this opinion, Wright has on this appeal essentially conceded the validity of this particular finding and has concentrated his argument here on the proposition that the unequal training given him in relation to that given Whitel-ock was the effective cause of his lack of qualification for promotion, and that being racially motivated, this constituted illegal discrimination.
At the outset, we think there is a serious question whether the evidence established a prima facie case in respect of the training opportunities allegedly denied. The McDonnell Douglas test for establishment of a prima facie case must certainly be adaptable to claims of unequal training opportunities,15 see Long v. Ford Motor Co., 496 F.2d 500, 505 & n. 11 (6th Cir. 1974), but this presents some difficulty where, as here, the training program is confined to a small group and is designed to qualify its participants for responsible managerial positions. Presumably all must necessarily be found “qualified” for the training program in general by virtue of their initial acceptance into it; but this cannot be thought necessarily to establish the qualification of each for every discrete assignment and special training opportunity that may be available from time to time within the program’s course. So to hold would in effect force all such training programs into a rigid lockstep configuration that could not take into account different incoming strengths and weaknesses, or differing rates of progress within a program. Particularly where the training program is inspired by affirmative action considerations, this could actually operate to thwart its goals. Be that as it may, even if Wright’s proof of his general qualification for the training program were deemed presumptively sufficient to round out proof of his prima facie case, he could not prevail on the total evidence, because of failures at later stages of the proof scheme.
As to each of the specific training opportunities given Whitelock and denied Wright, we agree with the district court that the defendants came forward with a legitimate nondiscriminatory reason that sufficed to dispel the adverse inference of discrimination generated by proof of the prima facie case. As to the admittedly superior supervision given Whitelock by the several supervisors, particularly Fraley, Whitelock’s own testimony that he more assiduously sought out supervisory assistance, and Fraley’s that she assumed White-lock was to be in her department following the training program16 stand as sufficient nondiscriminatory reasons, whatever they may say about the overall quality of the program’s design and administration. As to the more challenging assignments and special training opportunities given White-lock, the simple explanation advanced by defendants that they gave those assignments and opportunities for which the several trainees’ incoming abilities and progress equipped them at the time similarly provide a nondiscriminatory reason sufficient in the proof scheme to dispel the inferential proof of discriminatory motive. To deny to such an explanation even that tentative legal significance in a disparate training claim case would be simply to force all training programs having any minority group participants into a rigid pattern unadaptable to special needs and strengths of any of the participants. This cannot be thought commanded by Title VII.
The most critical denied opportunity as perceived and argued by Wright on this appeal was that to receive the benefit of work on the Special Projects Staff for as long a period as did Whitelock. Again, the [716]*716explanation offered by defendants for Whitelock’s earlier assignment, though more detailed, is essentially a simple one; and it too suffices as a legitimate nondiscriminatory reason at this stage of the proof scheme. In effect it was that for budgetary reasons only one of the members of this trainee class could be assigned at the time; that the assignment was made not merely to serve training purposes but to meet substantive needs of a hard pressed staff; and that of the three trainees then in the program, Whitelock was substantially better qualified to do the required work with the minimum of supervision that would be available. Again, whatever this may reveal about the lack of wisdom and perhaps sensitivity from a managerial standpoint of risking practically inevitable misunderstanding by providing special training opportunities to some but not all persons in such a program, it clearly passes muster as a nondiscriminatory reason not pretextual on its face.
We agree with the district court’s assessment that with respect to all the charged acts of discrimination, the defendants had come forward with sufficient evidence of nondiscriminatory reasons to dispel the prima facie showing. At this stage, consideration therefore properly turned for the district court, and turns for this court in review, to the question whether the evidence considered as a whole established that the nondiscriminatory reasons advanced were in fact mere pretexts for racially discriminatory treatment. On this the burden of persuasion was upon the claimant. We conclude with the district court that the evidence does not show the reasons to have been pretextual, and that accordingly Wright’s claim was not established.
At this stage, a claimant’s required proof of discriminatory motive has in effect been narrowed and focused upon the specific reasons advanced by the employer. The claimant need now only attack these, but the underlying requirement remains to prove that the real as opposed to now specifically “articulated” reasons was a racially inspired intent to treat less favorably. On this issue, the evidence, both that of the claimant and the employer, may take a variety of forms, but since the issue is fundamentally one of a state of mind, the available evidence will typically, as here, be wholly circumstantial. See McDonnell Douglas Corp. v. Green, 411 U.S. at 804-05, 97 S.Ct. 324. Of obvious importance is any evidence that tends to indicate the employer’s general conduct, hence likely motivation, in related treatment of the claimant and comparably situated members of the claimant’s protected group. See Furnco Construction Corp. v. Waters, 438 U.S. at 580, 98 S.Ct. 2943; McDonnell Douglas Corp. v. Green, 411 U.S. at 804-05, 97 S.Ct. 324. In the instant case, aside from the evidence offered of Fraley’s possibly specific racial bias,17 the evidence of the defendants’ general conduct in respect of Wright and other black trainees in the program tended overwhelmingly to establish a generally preferential rather than invidiously discriminatory attitude on the defendants’ part. That it was characterized by a considerable amount of fumbling, and was perhaps fundamentally misguided in relation to its affirmative action goals in many respects,18 cannot dispel this. The district court rightly pointed to this evidence as critical on the issue. We summarize it here. At the very outset, entrance requirements were lowered to encourage admission of blacks to the program. Thereafter blacks were actively recruited while whites were not. When the black trainees objected to the attitude of the defendant Fraley, Fraley was removed. When seating patterns caused concern, they were changed to meet the black trainees’ objections. [717]*717When the black trainees objected to White-lock’s assignment to the Special Projects Staff they were assigned to it, albeit later.19 Of the two black trainees who remained in the program to its stated end one, Sistare, was promoted despite an initial evaluation that he was not qualified, while the other, claimant here, was offered an extension of training in order to see if he might later qualify.20 There is simply.no countervailing evidence that tends to show any general racial bias at work against Wright and the other black trainees in this program. In the absence of more direct evidence of impermissible motivation for the specific acts charged, the circumstantial evidence of general attitude clearly fails to establish defendants’ stated reasons as pretextual. Indeed it tends to corroborate their nondiscriminatory motivation.
Perhaps in recognition of the force of this evidence of general motivation, Wright seems to argue here that defendants’ articulated reasons are shown to be pretextual simply by the fact that they do not indicate a purpose in each instance specifically to maximize Wright’s opportunities in relation to Whitelock’s. Thus, it is apparently suggested that with the ability to assign only one trainee to the Special Projects Staff, defendants’ only permissible, i. e., nonpretextual, basis for decision was deliberately to favor a black trainee over the sole white trainee without regard to any other considerations. However that may commend itself as a basis for action on principles other than the legal one we enforce, and however legally defensible it might be if done voluntarily, cf. United Steelworkers of America v. Weber, - U.S. -, 99 S.Ct. 2721, 61 L.Ed.2d 480 (1979), it clearly is not commanded by Title VII. A similar contention with respect to proof adequate to overcome articulated reasons for hiring decisions was advanced and specifically rejected in Furnco. See 438 U.S. at 576-78, 98 S.Ct. 2943. In doing so the Furnco Court pointed out that “Title VII prohibits . . having as a goal a work force selected by any proscribed discriminatory practice, but it does not impose a duty to adopt a hiring procedure that maximizes hiring of minority employees.” Id. at 577-78,98 S.Ct. at 2950. Transposing that principle to the training program decisions involved here, Title VII forbade making those decisions on a proscribed discriminatory basis; but it did not require making them solely on the basis of favoring a minority employee without regard to other considerations. To apply the disparate treatment proof scheme in this case as claimant would have us do would therefore [718]*718be to impose upon employers a burden not imposed by Title VII. This we may not do. We conclude with the district judge that on the evidence presented, no violation of Title VII was established.
Ill
In conclusion, it may be in order to step back from the necessarily technical discussion of “disparate impact” and “disparate treatment” legal doctrine to take a broader view of the issues and their resolution. The underlying grievance here plainly grew out of a positively motivated intention to better the employment conditions of minority employees. In the event, however, an affirmative action training program whose bona tides in creation has never been challenged went sour. The result was frustration and distress for all parties and eventually charges of racial discrimination in the administration of a program designed in major part to rectify past discrimination.
There are hard lessons to be learned. Any affirmative action effort worth its salt must involve some risks — for employers as well as employees. These risks must obviously be made reasonably tolerable for both if the efforts are to be continued with determination by both. Too high risks for employees in the form of unrealistic expectations can be expected to yield just the sort of frustration and antagonisms that occurred here. Ultimately this must result in chilling the willingness of minority employees to embark upon the risks at all. Employers have a clear moral and prudential obligation not to make those risks too high. At some point this obligation assumes legal proportions. The risks cannot be such that they constitute “built-in headwinds” for a protected group in relation to others so that the consequences are a provable “disparate impact” upon the former, or such that specific incidents of disparate treatment necessarily result from their imposition.
On the other hand, too high risks for employers in the form of legal requirements that all such programs undertaken must in effect be made to “work” for all members of the protected group can be expected to chill the zeal of employers voluntarily to undertake truly significant efforts.
Within controlling legal principles we have concluded that the risks inherent in the training program here could not fairly be said to have resulted in a “disparate impact” upon the protected group of which the individual claimant was a member, and that the various specific ways in which they were experienced by him could not be found racially motivated, hence to have involved “disparate treatment.” This is not to say that they may not have been most improvidently imposed. But improvidence, and even insensitivity, are not, standing alone, violations of Title VII. In the final analysis, as we have considered the evidence presented here, there is no way that Wright’s legal claim could be found established except by wrenching controlling principle to impose upon the employer an obligation to make this perhaps ill-conceived undertaking work for Wright. Such a rule given general application would impose so high a risk upon employers that their impulse to undertake any but the most innocuous and limited affirmative action efforts would inevitably be chilled. Such a consequence could not serve the fundamental purposes of Title VII.
AFFIRMED.