This is a 42 U.S.C. § 1983 action in which after two trials Henry Spell was awarded substantial damages against the City of Fayetteville, North Carolina (the City), and Charles McDaniel, a City police officer, as a result of physical injury inflicted on Spell by McDaniel while Spell was in McDaniel’s custody following Spell’s arrest. McDaniel and the City have appealed, assigning various trial rulings as error and challenging as unreasonable the amount of attorney fees awarded to Spell as prevailing party.
We find no reversible error in the trials and therefore affirm the judgment on the merits against McDaniel and the City. Except for its inclusion of a “contingency multiplier,” we also affirm the district court’s award of attorney fees.
I
Spell, admittedly inebriated on alcohol and quaaludes, was stopped by Officer McDaniel while driving an automobile in the City of Fayetteville. After talking with Spell and finding a quantity of quaaludes in his automobile, McDaniel arrested him *1384along with a passenger in Spell’s automobile, handcuffed the two of them and took them in a patrol car to the police station. There Spell was subjected to various sobriety tests, including a breathalyzer test, and was formally charged with driving while impaired and with the possession of quaa-ludes.1 Just after Spell completed the breathalyzer test and was returned, still handcuffed and inebriated, to McDaniel's direct custody, McDaniel, possibly angered by Spell’s failure to respond to his questioning, and in any event without any physical provocation, brutally assaulted Spell. When Spell warded off a blow toward his head by raising his arms, McDaniel seized his handcuffed arms, pulled them down and violently kneed Spell in the groin. The blow to Spell’s groin ruptured one of his testicles, necessitating its surgical removal. This resulted in irreversible sterility and of course in considerable associated pain and suffering.2
Spell then brought this § 1983 action naming as defendants McDaniel, the City of Fayetteville, the City Manager, the City Chief of Police, the Director of the police department’s Internal Affairs Division and two police department command sergeants. He structured the action as one against McDaniel in his individual and official capacities; against the City Manager, Smith, the Police Chief, Dixon, the Internal Affairs Division Director, Johnson, and the two command sergeants, Dalton and Holman, in their several official capacities; and against the City as a suable municipal corporation.
His pleaded theory of recovery against McDaniel individually was that McDaniel, acting under color of state law, had deprived him of rights secured by the fourth, fifth and fourteenth amendments by using excessive physical force against him in a custodial situation, thereby inflicting serious personal injuries.3 For this conduct he sought recovery of money damages against McDaniel in his individual capacity.
His pleaded theory of recovery against the City of Fayetteville was that the City was liable for damages under the doctrine of Monell v. Department of Social Services of the City of New York, 436 U.S. 658, 98 S.Ct. 2018, 56 L.Ed.2d 611 (1978), for the constitutional deprivation with consequent physical injuries directly inflicted by its employee McDaniel, because McDaniel’s conduct was pursuant to a municipal “policy or custom.’’ Id. at 694, 98 S.Ct. at 2037.
McDaniel denied inflicting any injury on Spell as a defense to the individual-capacity claim against him. The City also denied (for lack of sufficient knowledge or information) that McDaniel had inflicted injury on Spell, and alternatively denied that there *1385was any basis for imposing municipal liability upon it under Monell.4
The case then went to trial before a jury on the issues whether, as a matter of fact, McDaniel had assaulted Spell and was therefore liable individually, and if so, whether there existed a basis in law and fact for also imposing joint liability upon the City for the resulting constitutional deprivation.
After an 18-day trial, the jury returned verdicts finding McDaniel, in his individual capacity, and the City in its municipal capacity, jointly and severally liable. It awarded $1,000 in compensatory damages, and declined to award the punitive damages sought against McDaniel.
Spell then moved to set aside the $1,000 compensatory award as inadequate and for a new trial on the compensatory damages issue alone. The district court granted this motion, denying, inter alia, the defendants’ counter-motion for new trial on all the issues if the verdict was to be set aside for inadequacy of the damage award.
On the re-trial of the damages issue, the jury returned a verdict for $900,000 compensatory damages. After denying defendants’ renewed post-verdict motions for judgment n.o.v. or, alternatively, a new trial, the district court awarded Spell attorney fees and costs totalling $335,942.57. 616 F.Supp. 1069. Joint and several judgments against McDaniel and the City were then entered on the damage and fee awards.
This appeal followed.
Before us, the defendants join in contending that the district court erred in ordering a new trial on damages alone after setting aside as inadequate the first jury verdict; in making various evidentiary rulings; and in awarding unreasonably excessive attorney fees to Spell as prevailing party under 42 U.S.C. § 1988.
The City alone contends that the evidence did not warrant submission of the municipal liability issue to the jury and, in the alternative, that the court’s jury instructions on that issue were prejudicially erroneous.
Because they are the most serious and difficult, we consider first the municipal liability issues, then the others in order.
II
The City’s challenge to the submission of the municipal liability issue and the jury instructions upon it are interrelated. Both essentially question the district court’s understanding, hence application, of the law of municipal liability under § 1983, as it relates to incidents of police brutality such as that charged here.
We therefore begin by discussing that body of law as it applies to this case.
A
The Basic Principle
Municipalities are not liable under respondeat superior principles for all constitutional violations of their employees simply because of the employment relationship. Monell, 436 U.S. at 692-94, 98 S.Ct. at 2036-37. Instead, municipal liability results only “when execution of a government’s policy or custom, whether made by its lawmakers or by those whose edicts or acts may fairly be said to represent official policy, inflicts the injury.” Id. at 694, 98 S.Ct. at 2037-38.
Municipal “Policy” and “Policymaking”
While municipal “policy” is found most obviously in municipal ordinances, regulations and the like which directly command or authorize constitutional violations, see, e.g., Monell, id., at 661, 694, 98 S.Ct. at 2020, 2037 (official pregnancy leave policy) it may also be found in formal or informal ad hoc “policy” choices or decisions of municipal officials authorized to make and implement municipal policy, see Pembaur v. *1386City of Cincinnati, 475 U.S. 469, 106 S.Ct. 1292, 1301, 89 L.Ed.2d 452 (1986) (single “policy” decision by county prosecutor).
“Policy” in this context implies most obviously and narrowly a “course of action consciously chosen from among various alternatives” respecting basic governmental functions, as opposed to episodic exercises of discretion in the operational details of government. See City of Oklahoma City v. Tuttle, 471 U.S. 808, 823, 105 S.Ct. 2427, 2436, 85 L.Ed.2d 791 (1985); see also Pembaur, 106 S.Ct. at 1299 & n. 9. Correspondingly, “policymaking authority” implies authority to set and implement general goals and programs of municipal government, as opposed to discretionary authority in purely operational aspects of government. See Pembaur, 106 S.Ct. at 1299-1300 & n. 12; Bennett v. Slidell, 728 F.2d 762, 769 (5th Cir.1984) (en banc) (“policymaking authority is more than discretion ... more than the final say-so, as a matter of practice, on what water main will be replaced today”).5
While “finality” is a necessary attribute of “policymaking authority,” the possession of “final authority” does not alone constitute a municipal official a “policymaker.” The most critical factor is not the practical finality of an official’s “acts and edicts,” but their “policy” nature. See Pembaur, 106 S.Ct. at 1299-1300 (responsibility for establishing policy the key); Bennett, 728 F.2d at 769 (“policymakers ... decide the goals for a particular city function and devise the means for achieving those goals”). On the other hand, a municipal agency or official may have final authority to make and implement policy despite a municipality’s retention of powers of ultimate control over both policy and policymaker. The question is one of authority-in-fact. A municipal governing body may not avoid attribution of policy to itself simply by officially retaining unexer-cised ultimate authority to countermand a policy or to discipline or discharge the policymaker. See Bowen v. Watkins, 669 F.2d 979, 989-90 (5th Cir.1982); see also Bennett, 728 F.2d at 769 (policymaking authority exists despite retention by governing body of “prerogative of the purse and final legal control [to] limit or revoke the authority”).
Municipal “Custom” and its Creation
“Official policy” in the relatively narrow sense of discrete, consciously chosen courses of action by “policymakers” is not the only basis for imposing municipal liability. “Custom, or usage,” in the exact language of § 1983, may also serve. Monell, 436 U.S. at 690-91, 98 S.Ct. at 2035-36.6 And the existence of such a “custom or usage” may be found in “persistent and widespread ... practices of [municipal] officials [which] [although not authorized by written law, [are] so permanent and well-settled as to [have] the force of law.” Id. at 691, 98 S.Ct. at 2036 (citing and quoting Adickes v. S.H. Kress & Co., 398 U.S. 144, 167-68, 90 S.Ct. 1598, 1613-14, 26 L.Ed.2d 142 (1970)).
Attribution and Causation as Essentials of Municipal Fault
Because municipal liability results only when the municipality itself can be *1387directly charged with fault7 for a constitutional violation, it results only when policy or custom as above defined is (1) fairly attributable to the municipality as its “own,” Monell, 436 U.S. at 683, 98 S.Ct. at 2032, and is (2) the “moving force” behind the particular constitutional violation. Polk County v. Dodson, 454 U.S. 312, 326, 102 S.Ct. 445, 454, 70 L.Ed.2d 509 (1981).
Attribution of “Policy”
Policy, in the narrow sense of discrete, consciously adopted courses of governmental action may be fairly attributed to a municipality either because (1) it is directly “made by its lawmakers,” i.e., its governing body, Monell, 436 U.S. at 694, 98 S.Ct. at 2037, or (2) it is made by a municipal agency, see, e.g., id. at 661 & n. 2, 98 S.Ct. at 2020 & n. 2 (policy of city board of education and department of social services) or official, see Pembaur, 106 S.Ct. at 1300-01 (policy decision of county prosecutor), having final authority to establish and implement the relevant policy. A municipal agency or official may have final policy-making authority by direct delegation from the municipal lawmaking body, see, e.g., Wellington v. Daniels, 717 F.2d 932, 936 (4th Cir.1983) (delegation of law enforcement policymaking authority to police chief assumed) or by conferral from higher authority, see, e.g., Pembaur, 106 S.Ct. at 1301 (county prosecutor’s authority to act for county conferred by state law). Delegation may be express, as by a formal job-description, see Bennett, 728 F.2d at 769, or implied from a continued course of knowing acquiescence by the governing body in the exercise of policymaking authority by an agency or official, see id. (delegation “by conduct or practice [which] encourage[s] or acknowledge^] the agent in a policymaking role”).
Attribution of “Custom”
“Custom and usage,” in the sense of “persistent and widespread ... practices” by municipal agents and employees, may be attributed to a municipality when the duration and frequency of the practices warrants a finding of either actual or constructive knowledge by the municipal governing body that the practices have become customary among its employees. See id. at 768. Actual knowledge may be evidenced by recorded reports to or discussions by a municipal governing body. Constructive knowledge may be evidenced by the fact that the practices have been so widespread or flagrant that in the proper exercise of its official responsibilities the governing body should have known of them. See id.
Similarly, where a municipal policymaker has actual or constructive knowledge of such a course of customary practices among employees subject to the policymaker’s delegated responsibility for oversight and supervision, the “custom or usage” may fairly be attributed to the municipality as its own. See id at 769 (“custom” may be attributed to municipality through policymaker who has acceded to it).
Causation
When a municipal “policy or custom” is itself unconstitutional, i.e., when it directly commands or authorizes constitutional violations, see, e.g., Monell, 436 U.S. at 661, 694-95, 98 S.Ct. at 2020, 2037-38 (pregnancy leave policy), the causal connection between policy and violation is manifest and does not require independent proof. See Tuttle, 471 U.S. at 822, 105 S.Ct. at 2435 (“no evidence ... needed [in such a case] other than a statement of the policy”). But a policy or custom that is not itself unconstitutional in this strict sense must be independently proven to have *1388caused the violation.8 Proof merely that such a policy or custom was “likely” to cause a particular violation is not sufficient; there must be proven at least an “affirmative link” between policy or custom and violation; in tort principle terms, the causal connection must be “proximate,” not merely “but-for” causation-in-fact. Id. at 823, 105 S.Ct. at 2436; id. at 833 n. 9, 105 S. Ct. at 2441 n. 9 (Brennan, J., dissenting); see also Wellington, 717 F.2d at 936 (policy must be shown to have given at least “tacit authorization" to unconstitutional employee conduct). Neither the existence of such a policy or custom nor the necessary causal connection can be established by proof alone of the single violation charged. Tuttle, 471 U.S. at 823, 824, 105 S.Ct. at 2436, 2437; Wellington, 717 F.2d at 937.
The Principles as Applied to Constitutional Violations by Police
Application of these general principles to claims of municipal liability for specific incidents of unconstitutional conduct by law enforcement officials has raised special conceptual problems for the courts. This undoubtedly reflects conflicting, fundamental policy concerns about the exposure of municipalities to monetary liability for this most fundamental of “constitutional torts” at the local government level.
On one hand, there is concern that actual municipal culpability in these matters should not be masked and final responsibility avoided by overly rigid interpretations and applications of the concepts of policy or custom, policymaking authority, and causation. See, e.g., Tuttle, 471 U.S. at 833 n. 8, 105 S.Ct. at 2441 n. 8 (Brennan, J., concurring) (rejecting “metaphysical distinction between policies that are themselves unconstitutional and those that cause constitutional violations”). On the other hand, there is concern that a back-door vicarious liability principle should not be developed by overly tolerant interpretations and applications of those same key concepts. See, e.g., id. at 821, 105 S.Ct. at 2435 (plurality opinion) (overly tolerant concepts of policy and causation would impose municipal liability “simply because the municipality hired one ‘bad apple’ ”).
Municipal policymakers, including governing bodies, may of course in theory directly authorize unconstitutional police conduct just as any other form. In practice, however, such authorizations — “policies unconstitutional in themselves” — are understandably rare, or at least rarely surface in litigation in this realm. “Official” statements of law enforcement policy almost inevitably will specifically condemn rather than condone uses of excessive force or other unconstitutional conduct by police. And where there is no official statement respecting specific police conduct, it will be difficult if not impossible to imply an official municipal policy directly authorizing conduct at odds with federal and state constitutions and laws. Cf. Pembaur, 106 S.Ct. at 1301 (White, J., concurring) (“Local law enforcement officers are expected to obey the law_”).
Typically, therefore, claims of municipal liability for specific constitutional violations *1389by police have had to seek municipal fault in other sources than direct authorizations by policymakers and the necessary causation between fault and violation in more attenuated connections than direct commands.
Two basic theories have emerged for imposing municipal liability in the more typical situation where fault and causation cannot be laid to a municipal policy “itself unconstitutional.” The principal theory locates fault in deficient programs of police training and supervision which are claimed to have resulted in constitutional violations by untrained or mis-trained police officers. A second theory, sometimes imprecisely subsumed within the first, locates fault in irresponsible failure by municipal policymakers to put a stop to or correct a widespread pattern of unconstitutional conduct by police officers of which the specific violation is simply an example.
Although we have implicitly recognized the general viability of these theories in the course of rejecting particular claims of municipal liability, we have not been required earlier to elaborate their full details;9 nor has the Supreme Court yet done so.10 This case requires that we now do so, guided primarily by those Supreme Court decisions that have most directly suggested the theories’ bounds.
At the outset, we accept that in appropriate circumstances either or both of these theories may appropriately be invoked to impose municipal liability under the basic Monell principles above summarized. But we also accept that the most relevant Supreme Court decisions now require that each of the theories be carefully controlled at critical points to avoid imposing by indirection a form of vicarious municipal liability flatly rejected by Monell. Those critical points are (1) identifying the specific “policy” or “custom”; (2) fairly attributing the policy and fault for its creation to the municipality; and (3) finding the necessary “affirmative link” between identified policy or custom and specific violation. See Tuttle, 471 U.S. at 822-24, 105 S.Ct. at 2435-37 (plurality opinion); id. at 827-33, 105 S.Ct. at 2438-41 (Brennan, J., concurring); see also, Kibbe, — U.S. at -, 107 S.Ct. at 1120-22 (O’Connor, J., dissenting from dismissal of writ).
On this basis, we conclude as follows respecting the two theories.
Deficient training as culpable municipal “policy”
The way in which a municipal police force is trained, including the design and implementation of training programs and the follow-up supervision of trainees, is necessarily a matter of “policy” within the meaning of Monell. To the extent a particular training policy is fairly attributable to a municipality, it is “official municipal policy.” To the extent such an official municipal policy has deficiencies resulting from municipal fault that then cause specific constitutional violations by deficiently trained police officers, the municipality is liable under 42 U.S.C. § 1983.
*1390Such a training policy is fairly attributed to a municipality when it is designed, implemented, and its trainees supervised by municipal officials to whom the municipal governing body has effectively delegated final authority so to act. Delegation to such policymakers may be by formal directive, such as a job description, or by informal acquiescence in a known, continued exercise of authority, or by both in combination. Policymaking authority with respect to police training, as generally, is final if it is effectively final; the mere fact that the governing body or some other official retains ultimate but unexercised authority over both police training policy and policymakers does not allow a municipality to disclaim the policy as its own.
Only those deficiencies in police training policies that result from policymaker fault of at least the degree of deliberate indifference to or reckless disregard for the constitutional rights of persons within police force jurisdiction can give rise to municipal liability; mere negligence on the part of policymakers is not sufficient.11 Subject to this fault requirement, training policy deficiencies for which municipal liability may be imposed include not only express authorizations of specific unconstitutional conduct, but tacit authorizations, and failures adequately to prohibit or discourage readily foreseeable conduct in light of known exigencies of police duty.
Finally, a sufficiently close causal link must be shown between potentially inculpating training deficiency or deficiencies and specific violation. This requires first that a specific deficiency rather than general laxness or ineffectiveness in training be shown. It then requires that the deficiency or deficiencies be such, given the manifest exigencies of police work, as to make occurrence of the specific violation a reasonable probability rather than a mere possibility. In common parlance, the specific deficiency or deficiencies must be such as to make the specific violation “almost bound to happen, sooner or later,” rather than merely “likely to happen in the long run.” See Patzner, 779 F.2d at 1367 (training so deficient that “police misconduct inevitably occurs”).
Unconstitutional police practices as municipal “custom or usage” by condonation
Without having been directly authorized, or tacitly encouraged, or inadequately trained in specific ways by responsible municipal policymakers, police officers, like other public employees, may fall into patterns of unconstitutional conduct in their encounters with suspects, arrestees, persons in custody and others involved in law enforcement situations. This may result from a variety of factors not sufficiently traceable in origin to any fault of municipal policymakers to warrant treating the conduct as a reflection of “municipal policy” in the Monell sense.
If these unconstitutional practices become sufficiently widespread, however, they may assume the quality of “custom or usage” which, per § 1983, has the force of state “law” for purposes of invoking the remedies provided by § 1983. Adickes, 398 U.S. at 167-68, 90 S.Ct. at 1613-14. Such a developed “custom or usage” may then become the basis of municipal liability, but only if its continued existence can be laid to the fault of municipal policymakers, and a sufficient causal connection between the “municipal custom and usage” and the specific violation can then be established.
*1391Municipal fault for allowing such a developed “custom or usage” to continue requires (1) actual or constructive knowledge of its existence by responsible policymakers, and (2) their failure, as a matter of specific intent or deliberate indifference, thereafter to correct or stop the practices. Constructive knowledge may be inferred from the widespread extent of the practices, general knowledge of their existence, manifest opportunities and official duty of responsible policymakers to be informed, or combinations of these. The inculpating knowledge, whether actual or constructive, may be either that of the municipal governing body itself, or of municipal officials having final policymaking authority in municipal law enforcement matters.
A sufficiently close causal link between such a known but uncorrected custom or usage and a specific violation is established if occurrence of the specific violation was made reasonably probable by permitted continuation of the custom. Again, as in the case of deficient training policy, failure to correct the known practices must be such as to make the specific violation “almost bound to happen, sooner or later,” rather than merely “likely to happen in the long run.”
Deficient training “policy” and condoned “custom and usage” as alternative theories for establishing municipal liability
Because the “deficient training policy” and “condoned custom” theories are simply alternative ways of establishing municipal fault, hence potential Monell liability, for specific constitutional violations by police officers, they may be asserted as alternative theories in particular litigation. Fed.R.Civ.P. 8(e)(2). If sufficiently supported by proof, they may then be submitted to a jury as alternative bases for liability, and noncumulative liability may be found on the basis of either or both. This simply reflects the fact that in particular cases the question whether the “moving force” behind a specific violation was a policy that originated with municipal policymakers, or a custom that policymakers did not originate but effectively condoned, or both, may be determinable only by litigation. The most critical substantive difference between the two theories is that municipal fault may be ascribed for the first specific violation that results from proven policy, Tuttle, 471 U.S. at 822, 105 S.Ct. at 2435; Pembaur, 106 S.Ct. at 1297 n. 6, while fault for a violation resulting from condoned custom can only be ascribed when a pattern of comparable practices has become actually or constructively known to responsible policymakers, Wellington, 717 F.2d at 936 (“knowledge [may] be imputed to the supervisory personnel” only on basis of “history of widespread abuse”). Closely related is the principle that just as proof of a single violation will not support the inference that the violation resulted from a municipal “policy” of deficient training, see Tuttle, 471 U.S. 808, 105 S.Ct. 2427, 85 L.Ed.2d 791, so it obviously cannot support an inference that the violation resulted from a municipally condoned custom of comparable practices, see Wellington, 717 F.2d at 936.
B
As above outlined, the substantive requirements for establishing municipal liability for police misconduct are stringent indeed. The critical Supreme Court decisions have imposed this stringency in a deliberate effort to avoid the indirect or inadvertent imposition of forms of vicarious liability rejected in Monell. The City here contends that the evidence was insufficient to meet those stern substantive requirements, and that the district court therefore erred in denying its motions for directed verdict and judgment n.o.v.
We disagree. We conclude instead that the evidence was sufficient to support the jury verdict imposing municipal liability on either deficient training policy or condoned custom or usage theories.
A good framework for assessing the sufficiency of the evidence is the claimant’s core allegations of municipal liability, for his proof followed his allegations with considerable faithfulness and in great factual detail.
*1392Specifically, Spell pleaded that at the time of McDaniel’s assault upon him the various city officials named as defendants, “jointly and severally,” had the legal “duty ... for [sic] establishing, enforcing, directing, supervising and controlling policies, customs, practices, usages, and procedures to be used by police officials” of the City, so that their “edicts and/or acts” might “fairly be said to represent the official policy” of the City; that acting within that “duty” these named officials had by various acts of omission and commission “foster[ed] and encourage[d] an atmosphere of lawlessness, anarchy, repression and a repetitive policy, custom and practice of aggressive, abusive, and assaultive behavior and procedures toward individual detainees and arrestees which [on the date of McDaniel’s assault on Spell] represented the policy, practice, custom, usage and procedure” of the City, and that McDaniel’s assault upon Spell was “in furtherance of ... the practice, custom and procedure” of the City.
In greater detail, the complaint alleged, inter alia, that these defendant officials,
with knowledge of repeated allegations of abusive and assaultive behavior toward ... detainees and arrestees by [City] police officers ... repeatedly ... fail[ed] to enforce established procedures to insure the safety of individual arres-tees and detainees; ... established] and enforce[d] quota systems for arrests and citations ...; ... fail[ed] to discipline ... police officers ... who had been found to have committed abusive and assaultive behavior toward ... detainees and arrestees; fail[ed] and refuse[d] to competently investigate allegations of abuse and assault ... by ... police officers ...; reprimanded] ... intimidate[d], demote[d], and fire[d] police officers and officials who reported acts of abuse of authority by other[s] ...; covered] up acts of misconduct and abuse of authority by individual police officers and officials; ... rewarded] and commend[ed] ... police officers who displayed aggressive, abusive and assaultive behavior towards ... detainees and arrestees; repeatedly ... fail[ed] to adequately train and educate police officers in the use of reasonable force and proper use of authority; repeatedly fail[ed] to adequately supervise the actions of police officers and officials under their control and supervision.
As indicated, plaintiff’s evidence substantially tracked these essential allegations. Though that evidence was of course disputed in many details by the defendants’ countering evidence, much of it was undisputed and that which was disputed was not made manifestly incredible or implausible by anything in the record. Under the settled standard for assessing the sufficiency of evidence to survive a motion for directed verdict or judgment n.o.v., see, e.g., Lovelace v. Sherwin-Williams, 681 F.2d 230 (4th Cir.1982), plaintiff’s evidence clearly sufficed to establish the essential elements of both theories of municipal liability above summarized.
Assuming the credibility of the most critical oral testimony adduced by plaintiff and assessing the total evidence in its most favorable light, there emerged a picture of a police department whose members were either positively encouraged by training and deliberate cover-up to engage in uses of excessive force of the very type charged to McDaniel; or were tacitly encouraged to continue self-developed practices of this type by the deliberate failure of responsible municipal officials to exercise discipline or corrective supervision to halt the widespread, known practices; or were encouraged or authorized by both in combination.
To establish the existence of such a municipal policy or custom as the effective cause of Spell’s injury, plaintiff relied principally upon the testimony of seven lay citizens of the City, eight present or former officers of the City police department, an assistant state district attorney, the former legal advisor to the police department and upon internal records of the City police department. This evidence was essentially concentrated upon a time period of three years preceding McDaniel’s assault on Spell.
*1393The lay witnesses testified to a number of observed or directly experienced acts of brutality by city police officers of the type charged to McDaniel. They further testified that in each instance they filed complaints with the police department which resulted in no corrective or punitive action.
Even more critical was the testimony of former and still serving members of the police department. Officer Livingston testified to various incidents involving excessive uses of force by City police officers which were not punished because of an effective “code of silence” within the department; that City police officers were specifically trained in the technique of kneeing or grabbing testicles in order to subdue arrestees or persons in custody; that he had been concerned about the prevalence of physical violence within the department and had never heard of anyone being disciplined for such violence.
Officer Luskis testified to an observed pattern and specific incidents of brutality and excessive uses of force by members of the department; that he had seen new officers being trained in the technique of kneeing persons in the groin and that McDaniel’s kneeing of Spell was consistent with the training that he had received in the department; that there was little reason to report observed acts of brutality because his supervisor had told him that there was nothing that could be done about it.
Former officer Gardner testified that uses of excessive, unwarranted force were condoned by her supervisors, and that when she complained she was forbidden by her superior to interrupt or complain again about such incidents.
Officer Pirro, a twenty-five year veteran of the department, testified that defendant Dixon’s philosophy as chief of police was to “kick ass and take names”;12 that Dixon wanted “supercops” tougher than the soldiers at nearby Ft. Bragg; and that to further this goal, Dixon set up his own police academy to screen out women and nonaggressive male candidates.
Officer Fischer testified that Chief Dixon and defendant Holman attempted to project “tough guy” images in their training of recruits; that in his opinion they condoned the use of excessive force on arrestees; and he corroborated Officer Livingston’s testimony that there was an effective “code of silence” within the department which covered up such incidents.
Former Officer Matthews identified two former police officers rehired by Chief Dixon who Dixon knew to be overly aggressive, characterizing one as “one of the most brutal men I have ever seen.”
Douglas Canders, the former legal advis- or to the police department, testified that disciplinary procedures within the department had broken down to the point of being disregarded, in many instances by direction of Chief Dixon; that Chief Dixon advocated the use of excessive force by police officers in dealing with arrestees, and believed that corrective action should not be taken to punish or prevent its continuation; that he, Canders, left the department because of Chief Dixon’s attitude favoring the use of excessive force in police encounters.
James Ammons testified that while serving as an assistant state district attorney he had occasion to investigate and successfully prosecute two members of the police department for physical assaults upon suspects, one occurring upon the steps and in the lobby of the department’s center; and that in his opinion, based upon his personal observations and conversations with police officers, “the police department condoned and encouraged officers to use force and intimidation first, then to find out what the problem — what the situation — was” later.
McDaniel himself testified in deposition that he had been taught in the Fayetteville Police Academy, set up by Chief Dixon and run by Dixon and Sergeant Holman, that striking or grabbing the testicles of an *1394arrestee was an accepted method of subduing him.
Among the specific acts of police brutality recounted by these witnesses were the beating of a 14-year old boy by the defendant, Sergeant Dalton, while the child lay handcuffed on the ground in the presence of his mother and grandmother; the kneeing of a man’s testicles while the man was being attended by medical personnel; and pulling suspects from an automobile and beating them in the head and face while arresting and handcuffing them.
Finally, the plaintiff introduced into evidence police department records that graphically corroborated the testimonial evidence that specific instances of police brutality during the relevant time period were frequent but that complaints about them were consistently dismissed or disregarded, frequently with but cursory investigation. Referring to this evidence, the district court noted — with complete support in the record — that “the instances of confirmed and uncontradicted misconduct demonstrated in Defendant’s own internal affairs documents was ... devastating to defendants’ case.”
Evidence locating responsibility for this state of affairs was equally compelling in support of plaintiff’s claim of municipal liability. Most critically, the official job description for the position of Police Chief during the relevant period described it in general as “directing the full activities of the Police Department.” In more detail, it specified, inter alia, that the Police Chief “plans, programs, directs and evaluates the operation of the Police Department, ... formulates and implements police policy, procedures, rules, regulations and programs; ... promotes assigns and disciplines all personnel.” That this job description accurately described the authority in fact exercised by Chief Dixon during the relevant time period was indisputably corroborated by other evidence. Most critical perhaps was the testimony of defendants' own expert in police administration. In responding to the specific question, “Who makes policy for a police department?” he first defined “policy” as the “broad overall generalities or guidelines of where a department is going,” and then opined that
[pjolicy is established by the chief of police, and in this instance ... with the concurrence or acquiescence of the city manager, perhaps. But the chief of police has the responsibility for overall guidance — the policy. Procedure, which is how that policy is implemented, is the responsibility of the chief of police to develop in concert with his command officers.
Other testimony indisputably established that Chief Dixon fully assumed and acted upon the authority formally defined in this job description as verified by defendants’ expert witness. There was evidence that Dixon set up the Police Training Academy in which McDaniel, with others, received training; that he effectively set the tone of that training program, participating directly in its operation with Command Sergeant Holman, to whom he had delegated the requisite authority. There was further evidence, earlier noted, that in exercising his general authority over matters of supervision and discipline of police officers, he established a pattern of condonation, coverup, and disregard of reported police misconduct of the specific type charged to McDaniel.
It was conceded that Dixon as Police Chief was formally subject in all of this to the ultimate authority of the City Manager, and of course to the City’s governing body. But there was no evidence that either of these ever sought before the date of Spell’s injury effectively to curb or correct the way in which Dixon, aided by his command level subordinates, was in fact exercising his formal authority to train and supervise members of the police department as a critical aspect of his general authority to “direct[j the full activities of the Police Department.”
Finally, the evidence revealed that Dixon’s general authority in these matters was necessarily shared, both in the setting and the implementing of policies and programs, with subordinate officials, including the three Command Sergeants, Dalton, Holman, and Johnson, who were joined with *1395Dixon as official capacity defendants. Again, official job descriptions indicated that such a sub-delegation of Dixon’s officially delegated responsibilities was specifically authorized by the City as a contemplated aspect of his general authority as Police Chief.
This is but a general summary of the most critical evidence in a voluminous trial record related to the question of municipal liability. Viewed in its totality and under the appropriate standard, it amply supported submission of the claim of municipal liability to the jury, and the resulting jury verdict imposing liability.
More specifically, the evidence was sufficient to support findings by a properly instructed jury that at the critical time in issue, by formal delegation of the City’s governing body, Police Chief Dixon was an authorized policymaker for the City of Fay-etteville in matters of law enforcement, including specifically the training and supervision of members of the City Police Department in matters pertaining to physical encounters with arrestees and persons in lawful custody; that acting pursuant to that policymaking authority, Chief Dixon, in concert with subordinate command officers to whom he had duly delegated the requisite authority, had established and implemented a training program that was specifically deficient in its express or tacit encouragement of uses of excessive physical force in such encounters; that this deficiency resulted from the deliberate indifference of Chief Dixon and his authorized subordinates to the constitutional rights of persons within the City’s jurisdiction who might come into contact with deficiently trained police officers; that as a direct result of the deficient training which he had received under this training policy, defendant McDaniel brutally assaulted and injured plaintiff Spell.
Alternatively, or additionally, the evidence was sufficient to support findings by a properly instructed jury that at the critical time widespread uses of excessive force by members of the City’s police department had occurred and were known to Police Chief Dixon and his duly authorized subordinates who shared authority and responsibility for supervision and discipline of police officers engaged in such conduct; that Chief Dixon and his authorized subordinates however condoned such practices by deliberate cover-up, disregard and positive encouragement, causing the practices to continue as an established custom or usage; that this condonation resulted from the deliberate indifference of Chief Dixon and his authorized subordinates to the constitutional rights of persons within the City’s jurisdiction who might have encounters with police officers engaging in such practices; and that in specific accord with and in furtherance of this condoned custom or usage, McDaniel brutally assaulted and injured Spell.
C
Closely related to its argument that the evidence was insufficient to support a finding of municipal liability under Monell principles rightly interpreted, is the City’s argument that the district court committed a number of reversible errors in its jury instructions explaining and applying those principles. We disagree.
Tested rigorously against the con-cededly complex principles as they have now evolved, the instructions were understandably not precise in every critical detail, and certain portions if considered only in isolation might be thought technically erroneous. But construed overall and in light of the evidence and the specific legal theories upon which the claim and defense were tried, we find no error requiring reversal. The test of adequacy of instructions properly challenged on appeal is not one of technical accuracy in every detail. It is simply the practical one of whether the instructions construed as a whole, and in light of the whole record, adequately informed the jury of the controlling legal principles without misleading or confusing the jury to the prejudice of the objecting party. See generally 9 Wright & Miller, Federal Practice and Procedure: Civil, § 2558 (1971). Tested in this light, the instructions challenged on appeal pass muster; indeed are rather remarkable for their *1396overall accuracy and prescience in explaining what at the time was, and still remains, an evolving and difficult body of legal doctrine.
We note at the outset that there is a substantial question whether the City objected to the jury instructions at trial at the appropriate point and with sufficient specificity to preserve any right to challenge them on appeal. Fed.R.Civ.P. 51. We have decided nevertheless to consider their challenges, but in doing so we further note that the City’s formal assertions of error on appeal in turn suffer from considerable imprecision, so that fair review is made doubly difficult.
As we understand the City’s arguments in brief, the most critical of its challenges trace ultimately to problems associated with the way the lawsuit was structured as to parties-defendant. Discussion of the challenges requires a brief description of the party-structure problem. As noted in our opening recitation of the procedural history of the case, the plaintiff joined as named defendants on its municipal liability claim Smith, the City Manager; Dixon, the Chief of Police; Dalton, Holman, and Johnson, Police Department Sergeants; and the City itself. The named officials were joined originally and remained parties only in their respective official- capacities; the City was not originally joined but was added later by amendment.
While this ultimate structuring of parties defendant was not technically erroneous, it was not required in order to prosecute a claim of municipal liability to effective judgment against a municipality. Where, as here, no claim against officials in their individual capacities was made, a simpler, technically correct, and by far preferable structuring would have been to name the City as the sole defendant. See Kentucky v. Graham, 473 U.S. 159, 167 n. 14, 105 S.Ct. 3099, 3106 n. 14, 87 L.Ed.2d 114 (1985).13 In the event, the more cumbersome party structuring here produced the problems above noted. Specifically, it apparently led the district court to frame its instructions and its verdict submission in ways that focused jury consideration more directly than was ideal upon the culpability of each of the named defendant-officials as a possible basis for imposing municipal liability. Most critically, the court submitted a special verdict form that required the jury to find as to each of the named defendant-officials whether he had “developed” a “de facto” policy or failed to “nullify” a known “custom” that proximately caused McDaniel’s assault upon Spell. The jury dutifully responded by answering “Yes” as to each of the defendant-officials except City Manager Smith, as to whom it answered “No.”
Seizing upon this form of verdict submission in conjunction with its related instructions, the City urges two points.
The first is a narrow one that probably goes more fundamentally to the sufficiency of evidence to support submission of the issue in the form used than to the related jury instructions. Specifically, it is asserted that on the evidence it was not proper to allow the jury to consider that any of the defendant-officials other than City Manager Smith could have been a Monell “policymaker” in the matters at issue when Spell was injured. This was on the basis that undisputed documentary evidence established that until June 1, 1983, some five months before Spell’s injury, the City Civil Service Commission possessed final authority in all matters of police administration, and that after June 1, 1983, this final authority was reposed exclusively in City Manager Smith.14
*1397This contention fails for reasons earlier noted in Part II-A in our general discussion of municipal liability doctrine. If the agency and the city manager possessed exclusive authority in the critical matters here in issue, the evidence clearly established that it was only a paper, formal authority, never effectively exercised before Spell’s injury to curb or countermand the authority in fact being exercised by Dixon and his subordinates in the police department. See Bowen, 669 F.2d at 989. As indicated in our earlier discussion, municipal liability may not be avoided by such purely formal reservations of “final” authority. The district court therefore did not err in submitting issues that disregarded the asserted reservation here and allowed the jury to infer that final authority in fact lay in others as the de facto municipal policymakers.
The second point is broader. It asserts that in the court’s verdict submission and related instructions focussing on the culpability of particular defendant-officials, the court failed adequately to require that fault be traced directly to a policy or custom fairly attributable to the City as its own. In particular, the City asserts that the verdict form and related instructions permitted municipal liability to be imposed on the basis alone of findings of individual culpability of any of the named defendant-officials.
Analyzed in their totality and in light of the entire record, the instructions do not fail in this respect. The joinder of particular individuals as official-capacity defendants undoubtedly created an awkwardness for the court in explaining their status as parties not exposed to liability but whose conduct was critical to the imposition of municipal liability. But we are satisfied that the court adequately worked through the awkwardness and in the process sufficiently communicated the relevant principles of “policy” and “policymaking authority” as essential elements of municipal liability.15
The court distinctly pointed out at the outset of its instructions on the municipal liability claim that the City might only be found liable if the “unconstitutional action of its employee[ ] ... proximately resulted from ... policies, customs or practices of the defendant city, or an agency of the city, such as in this case, the Fayetteville Police Department” (emphasis added). Shortly thereafter the court instructed in classic Monell terms the meaning of “municipal policy,” pointing out that “policy” includes not only “official policy directives, regulations or ordinances” but “what the law considers de facto or in fact practices, customs or policies caused, maintained, tacitly encouraged or condoned by their supervisory personnel ... even though these ... have at no time received formal approval through the local government’s decision making channels.” Elsewhere the court had effectively instructed that “municipal policy or custom” as above defined could only be made by those “whose edicts, acts *1398or omissions ... may fairly be said to cause, maintain, tacitly encourage or condone the ... policy, custom, practice or procedure of the ... City.” Such persons, the court indicated, were “representative” of the City in the sense that “policy or custom” created or condoned by them was that of the City in legal contemplation. Hence, it was clear that when the court used the term “representative defendant” in its instructions and verdict submission it was describing a defendant whom the jury must have found to have the requisite authority to establish or condone municipal policy, i.e., to have final policymaking authority. The court’s various indications that the conduct of such a “representative defendant” was attributable to the City was therefore a legally correct statement of the critical attribution principle rather than an erroneous negation of the “policy” and “policymaking authority” elements of municipal liability.
The City also challenges the adequacy of the jury’s instructions on the causation element. This challenge is without merit.16 The court effectively instructed the jury at various points that in order to impose liability upon the City “through” the conduct of “one or more” of the “representative,” i.e., policymaking, defendants, the jury must find that the conduct had established a policy or condoned a custom that “proximately” caused McDaniel to injure Spell. This was further elaborated by the court’s explanation, drawing on related tort principles, that the culpable conduct of establishing policy or failing to “nullify” custom must be found to have “set in motion a series of acts by others which the representative defendant should have foreseen would cause his subordinates, including defendant McDaniel, to in fact inflict constitutional injury upon ... citizens.” This was a completely adequate instruction, considered in total context of the record, to communicate the requirement that an “affirmative link” between policy or custom and specific violation be found.
Finally, the City challenges the court’s instructions on the alternative theory of “condoned custom.” Essentially, the complaint is that the court did not adequately explain that only a known, widespread “custom or usage” of unconstitutional conduct could be attributed, by condonation of policymaking officials, to the City. Specifically, the City points to a passage from the instructions in which the court instructed that “the citizens of a municipality do not have to endure an overly widespread or unduly prolonged pattern of past misconduct before municipal liability can be found.” Again, we find no reversible error in this particular portion of the instructions.
The critical passage challenged simply followed up the court’s immediately preceding instruction that what must be found was “a general and persistent pattern in practice ... so common and well settled as to constitute a custom or pattern of official misconduct,” which was then elaborated by the critical admonition that “a single isolated act is insufficient to establish a de facto policy.” In this immediate context as well as in the more general context, this passage was therefore a quite acceptable counterpoise, using the qualifiers “overly” and “unduly,” to the basic proposition that only practices “general and persistent,” “common and well settled,” could constitute condoned custom or usage.
We therefore find no reversible error in the particular instructions sought to be challenged by the City.
For reasons deemed sufficiently important, we have reviewed the instructions more generally, following the rule in this circuit that despite the important strictures of Fed.R.Civ.P. 51, instructional error may be noticed in civil cases despite fail*1399ures to make proper contemporaneous objection, when the error is “plain” or “fundamental.” See, e.g., Furka v. Great Lakes Dredge & Dock Co., 755 F.2d 1085, 1089 (4th Cir.1985); Miller v. Premier Corp., 608 F.2d 973, 983 (4th Cir.1979). We are prepared to reverse on such a basis only when we can conclude that a particular jury instruction must necessarily have caused the jury to act in complete ignorance of, or to have misapplied, fundamentally controlling legal principles to the inevitable prejudice of an aggrieved party. Ordinarily, however, we will only disregard failures to object properly at trial if specific challenge is made, though belatedly, on appeal. Here, because of the complexity, and importance, and still evolving nature of the principles of municipal liability that control decision in this case, we have thought it appropriate to review sua sponte the instructions even more generally to determine if error that “fundamental” may have occurred. We conclude that it did not.
We are satisfied that the district court’s instructions adequately informed the jury of the essential principles of both the “deficient training policy” theory and the related, alternative “condoned custom or usage” theory of municipal liability and appropriately submitted them as alternative legal/factual theories. In particular, we have considered whether the instructions imposed sufficiently stringent substantive standards at the critical points of those alternative theories: identification of the specific policy or custom claimed to have caused the specific violation; fair attribution of the policy or custom and a requisite degree of fault to the City; establishment of a sufficiently close causal connection between policy or custom and specific violation.
The court’s instructions adequately informed the jury that it must find that police training “policy” gave at least “tacit encouragement” to the specific misconduct — excessive use of force — charged to McDaniel, or, alternatively, that there had developed a “well settled” custom or practice which included such misconduct. The instructions then required that such a specifically deficient training policy should have been established, or a specifically inclusive custom should have been knowingly condoned by “representatives” of the City duly authorized to establish policy, or to condone developed practices. The instructions required that the establishment of deficient training policy or condonation of culpable custom should be traceable to the “gross negligence” or “deliberate indifference” of those authorized and responsible to act for the City. Finally, as earlier indicated, the instructions properly required a finding of sufficiently close connection between policy or custom and specific violation.
Stepping back somewhat from the conceptual complexities and difficulties of evolving municipal liability doctrine and assessing the overall thrust of the court’s instructions in practical terms, we have no doubt that the essentials of the doctrine as it now stands were accurately conveyed. The jury was sufficiently informed that it could only impose liability on the City if it was persuaded that law enforcement officials to whom the City had effectively delegated authority to act for it in the policy matter of training and supervising its police had either intentionally or by gross negligence or deliberate indifference so mis-trained them or failed to supervise them, or both, that the police had been given reasonably to believe that they were generally free to use excessive force in any police-citizen encounters without fear of sanction, and that as a direct result of this informal “policy or custom” McDaniel had unjustifiably assaulted Spell. That suffices to uphold the instructions against both specific challenge by the City and sua sponte review by the court undertaken to protect against fundamental injustice in application of an evolving body of legal doctrine.
Ill
The City has made several other assignments of trial error. We have considered each of them and find none with merit. Only three warrant even limited discussion.
*1400The first challenges the district court’s order of a new trial confined to damages alone, following the first jury’s return of a verdict finding liability against both McDaniel and the City but awarding Spell only $1,000 in compensatory damages.
The City does not seriously challenge— nor could it — the court’s determination that the award was inadequate, but asserts that if new trial was warranted it should have been on all issues. We disagree.
The decision was a discretionary one, and it lay well within the bounds of informed discretion. Indeed, to have proceeded otherwise may well have been an abuse of discretion here. The controlling legal principle of course is that in such situations a new trial on all issues is indicated if the verdict “could only have been a sympathy or compromise verdict ... [b]ut where there is no substantial indication that the liability and damage issues are inextricably interwoven ... a second trial limited to damages is entirely proper.” Great Coastal Express, Inc. v. International Brotherhood of Teamsters, 511 F.2d 839, 846 (4th Cir.1975) (emphasis added). This principle simply reflects the common sense proposition that a verdict this manifestly inadequate to compensate for undisputed physical and economic harm is practically bound to reflect one of two things: either a compromise reflecting a jury’s desire to give at least a little something to an attractive claimant despite serious reservations or outright doubt about the proof of defendant’s culpability; or an unwillingness adequately to compensate an unattractive claimant despite overwhelming proof of culpability. Whether a particular verdict reflects the one or the other is quintessentially a decision committed to the informed discretion of the judge who has conducted the trial and can best estimate the relative possibilities. Here, the record completely supports the district court’s assessment of the proper course.
In a related assignment of error, the City contends that the . district court abused its discretion in declining to set aside the second jury’s compensatory award of $900,000 as excessive. Here again, of course, the district court’s ruling is a discretionary one, and indeed is one that we review with even more than ordinary deference. See Grunenthal v. Long Island Rail Road Co., 393 U.S. 156, 160, 89 S.Ct. 331, 334, 21 L.Ed.2d 309 (1968) (only to determine if “untoward, inordinate, unreasonable or outrageous”); Simmons v. Avisco, Local 713, Textile Workers Union, 350 F.2d 1012, 1020 (4th Cir.1965) (only to determine whether “not merely excessive but ‘monstrous’ ”).
Under this standard we cannot find error in the district court’s ruling; indeed it seems to us eminently sound. Although Spell’s medical expenses were relatively low ($2,041), his hospital stay short (four days), and his ability to function sexually not permanently impaired, the evidence showed that the assault caused him intense pain, that his damaged testicle enlarged five to seven times its normal size as a result, that it was like “a smashed piece of fruit” with the outer covering torn and the internal contents passing through the tear, that surgical removal of the testicle led to permanent disfigurement and that, on account of an earlier illness, the assault left Spell irreversibly sterile.
Finally, the City argues that the district court erred in admitting evidence of the City’s settlement of an earlier police brutality action brought by another party, and in admitting evidence of arrest quotas and of sexual discrimination and harassment within the police department towards arrestees and trainees.
The first item of evidence was properly admitted to show, as an essential element of Spell’s “condoned custom” theory of liability, that the City was sufficiently aware of the existence of a developed practice or custom of such conduct. Its relevance substantially outweighed any risk of unfair prejudice within contemplation of Fed.R. Evid. 403, a conclusion bolstered by the fact that the court gave a limiting instruction which properly confined jury consideration to this purpose.
The second type of evidence may have lain at the margins of relevance and unfair *1401prejudice, but we are satisfied that its admission was not an abuse of discretion, certainly not error requiring reversal on this record.
In his complaint Spell specifically alleged that the defendants “did establish and enforce quota systems for arrests and citations to encourage arrest and detention of citizens without regard for due process of law and probable cause” and “did foster and encourage an atmosphere of lawlessness, anarchy [and] repression ...” which reflected the policies or customs of the City of Fayetteville, and which resulted in the injury inflicted by McDaniel. Although, as defendants contend, “this case concerns an alleged assault by a white male police officer upon a white male arrestee,” in its Monell dimensions, Spell’s complaint raised the larger issue of police department customs of authorizing or condoning the use of excessive force against arrestees. The issue of municipal liability as framed by Spell’s complaint involved a broad inquiry into the training, supervision, investigation, and discipline of police officers. Evidence of the existence of arrest quotas, and discriminatory attitudes and practices which, according to Spell, fostered an atmosphere of aggressive and lawless behavior and compromised the uniformity and quality of police department training and supervision, thus bore directly upon the question of municipal liability for the assault on Spell. As the district court concluded, “evidence of defendants’ misconduct lay at the heart of plaintiff's complaint. It was not introduced to prove defendants’ character, rather it was introduced and admitted as substantive evidence in support of the Monnell [sic] allegations in plaintiff’s complaint.” Spell v. McDaniel, 604 F.Supp. at 649. We agree with this assessment.
IV
The City’s final assignment of error is to the district court’s award of attorney fees to plaintiff as a prevailing party under 42 U.S.C. § 1988.
The court awarded Spell a total of $325,-303.75 in attorney fees representing 1,964.5 hours billable at rates ranging from $100-125 per hour. A lodestar of $224,758.50 was enhanced by a “contingency” adjustment or “multiplier” of 1.5 to compensate Spell’s counsel for the initial risk of not prevailing and thus not collecting a fee. See Spell v. McDaniel, 616 F.Supp. 1069, 1107-10 (E.D.N.C.1985). The defendants contend that the district court erred by including within the lodestar amount hours spent by Spell’s attorneys on “non-legal" tasks such as serving subpoenas, photocopying documents, and filing pleadings; by allowing the $100-125 hourly rates requested by Spell’s counsel despite their failure to offer evidence of their customary rates; by setting hourly rates of $100 for attorneys Holt and Richardson although each had less than five year’s trial experience when this litigation commenced; by applying uniform rates to all services rendered regardless of the nature of the work performed; and by presuming the propriety of a contingency adjustment in § 1988 cases.
A
An allowance of attorney's fees by a district court, which has “close and intimate knowledge of the efforts expended and the value of the services rendered, must not be overturned unless it is 'clearly wrong.’ ” McManama v. Lukhard, 616 F.2d 727, 729 (4th Cir.1980) (quoting Barber v. Kimbrell’s, Inc., 577 F.2d 216, 226 (4th Cir.1978)) (citing Lea v. Cone Mills Corp., 467 F.2d 277, 279 (4th Cir.1972) (quoting United States v. Anglin & Stevenson, 145 F.2d 622, 630 (10th Cir.1944))). Applying this standard to the careful analysis of the district court, we conclude that the defendants’ contentions are without merit except as to the court’s inclusion of a “contingency multiplier.”
B
With respect to its inclusion within the lodestar amount of hours spent by Spell’s counsel on “non-legal” tasks, the district court observed that the defendants’ argument was “predicated on the erroneous assumption that a single, correct staffing pattern exists within every law firm for every *1402lawsuit.” 616 F.Supp. at 1096. Rather, as the court found, in some circumstances, it may be more efficient and reasonable for counsel to perform such tasks. Thus, for example, Spell’s attorneys provided affidavits indicating that in many cases their service of subpoenas was accompanied by a final interview with witnesses; that personal service was often of particular importance because many of Spell’s witnesses required constant reassurance; and that while photocopying documents, they screened, reviewed and arranged more than 14,000 pages of police department internal reports, “perhaps the most important time spent on this case by plaintiff's counsel.” Id. at 1096-97. As a precautionary measure, the court nevertheless deducted ten hours from those requested “[t]o the extent secretaries could have been competently employed to perform some of this work.” Id. at 1097.17
The § 1988 fee applicant bears the burden of establishing the reasonableness of the hourly rates requested. Specifically, the applicant must produce specific evidence of the “prevailing market rates in the relevant community” for the type of work for which he seeks an award. Blum v. Stenson, 465 U.S. 886, 895-96 & n. 11, 104 S.Ct. 1541, 1547 & n. 11, 79 L.Ed.2d 891 (1984); National Association of Concerned Veterans v. Secretary of Defense, 675 F.2d 1319, 1325 (D.C.Cir.1982). The prevailing market rate may be established through affidavits reciting the precise fees that counsel with similar qualifications have received in comparable cases; information concerning recent fee awards by courts in comparable cases; and specific evidence of counsel’s actual billing practice or other evidence of the actual rates which counsel can command in the market. National Association of Concerned Veterans, 675 F.2d at 1325-26; Ramos v. Lamm, 713 F.2d 546, 555 (10th Cir.1983).
The district court acknowledged that Spell’s counsel had “barely” made the required showing. Because they omitted information regarding their own billing practices and provided only a few generalized affidavits from local counsel, the district court was required to “substitute its personal knowledge and experiences in lieu of substantive evidence to a greater degree than it would have hoped.” 616 F.Supp. at 1102. The court noted, however, that the defendants did not object to the rates requested, and that the factors articulated in Johnson v. Georgia Highway Express, Inc., 488 F.2d 714, 717-19 (5th Cir.1974), which were embraced by this court in Barber, 571 F.2d at 22618 militate strongly in favor of compensating Spell’s counsel at the rates sought. Among other things, the court found that the litigation presented challenging issues of both fact and law of considerable complexity and magnitude; that Spell’s counsel incurred substantial opportunity costs in pursuing the litigation, given the drain of resources on their four-person firm and the unpopularity of their case within the community; and that the rates requested were in line with other *1403awards in recent years in similar cases.19 616 F.Supp. at 1103-05. Although, as the district court observed, the rates allowed were at the “apogee of the range of rates that prevail in this district for lawyers of comparable qualifications and experience who handle cases of this sort,” id. at 1105, we cannot say that they were “clearly wrong.”
Nor can we say that the district court abused its discretion in allowing attorneys Holt and Richardson to charge their time at $100 per hour although each had less than five year’s trial experience when this litigation commenced. Counsel’s experience is only one of many factors to be considered in determining a reasonable fee. See Johnson, 488 F.2d 714.
As to the district court’s application of uniform rates to all services rendered by Spell’s counsel without regard to the nature of the work performed, we likewise cannot say that the result was “clearly wrong.” As the district court noted, the lower courts have adopted varying approaches to the problem of valuing legal services, with some utilizing different rates for different types of work, particularly in-court and out-of-court or primary and secondary services, see e.g., Whitley v. Seibel, 676 F.2d 245, 253-54 (7th Cir.1982); Miles v. Sampson, 675 F.2d 5, 9 (1st Cir.1982); United Nuclear Corp. v. Cannon, 564 F.Supp. 581, 589 (D.R.I.1983); Langdon v. Drew Municipal Separate School District, 512 F.Supp. 1131, 1139 (N.D.Miss.1981), and others applying uniform rates for all services rendered, White v. City of Richmond, 713 F.2d 458, 461 (9th Cir.1983); Maceira v. Pagan, 698 F.2d 38, 40-41 (1st Cir.1983); Bonner v. Coughlin, 657 F.2d 931, 937 (7th Cir.1981). The district court applied flat rates in this case consistent with its own previous awards and with those of other judges within the district. 616 F.Supp. at 1103. The court found, moreover, that Spell’s “counsel’s knowledge, skills and abilities were as fully engaged in their out-of-court work as they were in-court.” Id.
We therefore find no error in the district court’s determination of the “lodestar fee.”
C
As indicated, the district court, having computed the “lodestar” fee figure, then enhanced it by a 50% “multiplier,” expressly to compensate counsel for the risk of not prevailing and therefore being paid nothing under their contingency fee contract.20 Defendants challenge this primarily on the basis that the district court acted on a “presumption” of entitlement to a contingency multiplier rather than by a careful exercise of its discretion.
We disagree with this specific basis of challenge. The district court properly recognized that the “presumption” ran the other way: that the lodestar figure presumptively represented a reasonable fee, with the burden on the fee claimant to establish the propriety of an upward contingency adjustment. See 616 F.Supp. at 1107 n. 59. Based upon this understanding, the district court indeed undertook an extended, meticulous analysis of those factors it considered relevant to the use of a contingency multiplier under then controlling legal precedent. Noting that the general propriety of contingency multipliers had been expressly reserved by the Supreme Court, see 616 F.Supp. at 1107 (citing Blum v. Stenson, 465 U.S. at 901 n. 17, 104 S.Ct. at 1550 n. 17), the court relied for its allowance of a multiplier on a number of lower federal court decisions which had assumed their general propriety. See 616 F.Supp. at 1108-10.21
*1404It would be difficult to fault the district court’s factual and legal analysis under the principles it understandably thought at the time controlled its discretionary allowance of the contingency multiplier. But the Supreme Court has in the interval now spoken to the question reserved in Blum, and in a way that we conclude makes a contingency multiplier inappropriate here as a matter of law.
In Pennsylvania v. Delaware Valley Citizens’ Council for Clean Air, — U.S. -, 107 S.Ct. 3078, 97 L.Ed.2d 585 (1987), a 5-4 majority of the Court held that a contingency multiplier may be permissible in appropriate cases,22 though another 5-4 majority thought the multiplier used in the case before it not warranted.
Within the divided Delaware Valley Court, Justice O’Connor’s position in concurrence controls on the circumstances in which a contingency multiplier may ever be allowable.23 Critical to that concurring position as it affects the instant case are the propositions that (1) “no enhancement for risk is appropriate unless the applicant can establish that without an adjustment for risk the prevailing party ‘would have faced substantial difficulties in finding counsel in the local or other relevant market,’ ” this being the basic purpose of § 1988, id. at -, 107 S.Ct. at 3089-92; and (2) no enhancement for risk should be allowed on the basis of “ ‘legal’ risks or risks peculiar to the case” since these risks are adequately taken into account in fixing the “lodestar” figure, id. at-, 107 S.Ct. at 3091.
Specifically applying those constraints, Justice O’Connor provided the majority necessary to reverse allowance of the multiplier in that case on the alternative ground that it was inappropriate on the facts of the case even if not generally inappropriate. This majority emphasized that there was no finding of record that an enhancement for risk was necessary in the relevant market to insure that “poor clients with good claims [could] secure competent help,” id. at-, 107 S.Ct. at 3089; and that the lower courts had instead found enhancement appropriate on the impermissible basis of the legal and practical risks peculiar to the particular case, rather than on the basis of any general difference in relevant market treatment of all comparable contingent fee cases, id. at-, 107 S.Ct. at 3089-92.
We conclude that these same critical constraints make the allowance of a contingency multiplier inappropriate in the instant case.
Though the district court noted, and we can assume, that the case was an “unpopular” one in the community, there is no finding of record that a contingency en*1405hancement was necessary to insure that competent counsel would nevertheless be willing to undertake contingency representation in such cases. More critically, we think that the record strongly suggests, if it does not compel, a directly opposite conclusion. In its lodestar analysis, the district court alluded to the fact that “there are a number of pending lawsuits in federal court against the [police department] and the City,” and that plaintiff's counsels’ time spent in conference with other counsel in those cases was for that reason practically justifiable. . 616 F.Supp. 1099. From this we think it obvious that plaintiff here could not possibly show that “without risk-enhancement, plaintiff would have faced substantial difficulties in finding counsel in the local or other relevant market,” id. — U.S. at -, 107 S.Ct. at 3089 (plurality opinion). The picture that emerges from the entire record is instead that these counsel and others in related cases had no general reluctance to undertake contingency representation in this type case notwithstanding any general “unpopularity” they might have among the local lay populace. Indeed, it would appear that there was no dearth of competent counsel willing to take this and comparable cases on a contingency basis as being “good plaintiffs’ cases” without any guarantee of special enhancement of reasonable “lodestar” fees.
Furthermore, here as did the lower courts in Delaware Valley, the district court essentially justified its allowance of a multiplier on the peculiar risks, complexities, and difficulties of the instant case rather than on any perception that enhancement was required to insure adequate representation in the run of like contingency fee cases. See 616 F.Supp. 1107-10. Specifically, the court, relying essentially on the mode of analysis followed by the Third Circuit in its pre-Blum decision in Lindy Brothers Builders v. American Radiator & Standard Sanitary Corp., 540 F.2d 102 (3d Cir.1976), looked primarily at the case-specific “burden” on the plaintiff in terms of the case’s complexity, dubiety of success on the merits, and difficulty of proving damages; the case-specific practical risks of uncompensated time and expense assumed by counsel in undertaking representation; and the delay in receiving payment, see 616 F.Supp. at 1107-10. As we read Delaware Valley, a majority of the Supreme Court considered that these case-specific risk factors may not properly be drawn upon to enhance a “lodestar” fee with a “contingency multiplier.” See Delaware Valley, — U.S. at -, 107 S.Ct. at 3079-87 (plurality opinion); id. at -, 107 S.Ct. at 3089 (concurring opinion).
Accordingly, we conclude that the “contingency multiplier” allowed in this case is not legally justified, and must be excised from the total fee award.
V
We therefore affirm the judgment against McDaniel and the City on the merits, and we vacate the district court’s order awarding costs and attorney fees for modification in conformity with this opinion.
AFFIRMED IN PART; VACATED AND REMANDED IN PART.
. Spell later pled guilty to the possession charge which was contained in a multi-count indictment that also charged two counts of narcotics trafficking for which he was convicted after trial. At the time of trial of this § 1983 action, he was serving a seven year sentence growing out of those convictions, a fact brought out to the jury in Spell’s own testimony on direct examination.
. These are the essential facts necessarily accepted in substance by the jury in finding McDaniel liable. They were disputed by McDaniel, who denied making any assault on Spell and speculated that the conceded injury resulted from a pre-arrest occurrence. Though acceptance of these facts required outright rejection of McDaniel's testimony and that of another officer circumstantially corroborating McDaniel’s, there was more than ample evidence supporting the critical finding. The district court, denying defendant’s motion for judgment n.o.v. and alternatively for new trial, expressed flat incredulity at the testimony offered to support McDaniel’s denial that he ever physically assaulted Spell. See Spell v. McDaniel, 604 F.Supp. 641, 647-48 (E.D.N.C.1985).
.Spell also asserted deprivation of rights under the Eighth Amendment, and the Privileges and Immunities Clause. The district court properly dismissed these. 616 F.Supp. 1069, 1076. The district court considered the claim under both substantive due process and fourth amendment theories, treating the two as essentially congruent. 591 F.Supp. 1090, 1101, 1103. In reviewing the case, we have considered the claim as one most appropriately invoking fourth amendment protections against unreasonable uses of force in effecting and maintaining arrests. See Tennessee v. Garner, 471 U.S. 1, 105 S.Ct. 1694, 85 L.Ed.2d 1 (1985); cf. City of Oklahoma City v. Tuttle, 471 U.S. 808, 816-17, 105 S.Ct. 2427, 2432-33, 85 L.Ed.2d 791 (1985). The defendants do not contest this as a viable basis of claim.
. The City (and the official capacity defendants) also raised as other "affirmative defenses” good faith immunity of the City and the officials, the existence of an adequate state remedy (presumably under Parratt v. Taylor, 451 U.S. 527, 101 S.Ct. 1908, 68 L.Ed.2d 420 (1981)) and, apparently, state law sovereign immunity for discretionary actions of the official capacity defendants. The district court properly declined to submit these as possible defenses to claims of fourth amendment and substantive due process violations.
. This and following specifically approving citations to Bennett v. Slidell reflect our agreement with much of the comprehensive analysis by the en banc majority in that case of the nature of "municipal policy,” "policymaking authority,” and related concepts of municipal liability under Monell. Our extensive agreement with that conceptual analysis implies no view on the result reached on the facts of that case by the closely divided en banc court.
. As Justice Stevens continues to point out, the term "policy" is not found in the text of § 1983, while "custom, or usage” is. See Pembaur, 106 S.Ct. at 1302 (Stevens, J., concurring); Tuttle, 471 U.S. at 841-42, 105 S.Ct. at 2445-46 (Stevens, J., dissenting). Neverthless, "policy" has acquired the separate and distinctive meaning described in text as a result of the Supreme Court’s municipal liability decisions in Monell and its progeny. That the Court recognizes "municipal policy” in the judicially developed sense and "custom, or usage" in the statutory sense as different legal concepts seems plain, though the Court has not since Monell had the occasion to discuss "custom, or usage" in the detail it has discussed "policy.” See Pembaur, 106 S.Ct. at 1299 n. 10 (distinction noted and viability of claims under both recognized).
. Though criticized by commentators, e.g., Mead, 42 U.S.C. § 1983 Municipal Liability: The Monell Sketch Becomes a Distorted Picture, 65 N.C.L.Rev. 517, 556-60 (1987), and still disputed by one member of the Supreme Court, see Tuttle, 471 U.S. at 834-44, 105 S.Ct. at 2441-47 (Stevens, J., dissenting), the settled Supreme Court position is that direct municipal fault is the essence of municipal liability under § 1983. Id. at 818, 105 S.Ct. at 2433 (plurality opinion) ("fault-based analysis” required); id. at 830-31, 105 S.Ct. at 2439-40 (Brennan, J., concurring) (absent independent evidence of policy or custom, "no way of knowing whether the city is at fault”).
. Some members of the Supreme Court have raised the question whether municipal liability may ever be imposed except for "policy” which is "itself unconstitutional," as, e.g., was that in Monell, or in Pembaur (by retroactive declaration). See Pembaur, 106 S.Ct. at 1301-02 (White, J., concurring) (where controlling law forbids particular conduct, municipal officials "cannot be said to have the authority to make contrary policy”); Tuttle, 471 U.S. at 824 n. 7, 105 S.Ct. at 2436 n. 7 (Rehnquist, C.J., plurality opinion) (question reserved). Despite these express reservations, we assume that a majority of the Court is not now prepared so drastically to limit the scope of municipal liability under § 1983. We consider this at least implicit in the Tuttle plurality's reservation of the issue whether, subject to proper jury instructions imposing strict proof requirements, liability might be imposed on the basis of a deficient training "policy." See id., at 823-24 & n. 7; 105 S.Ct. at 2436-37 & n. 7; see also City of Springfield, Massachusetts v. Kibbe, — U.S. -, -, -, 107 S.Ct. 1114, 1117, 1120, 94 L.Ed.2d 293 (1987) (O’Connor, J., dissenting from dismissal of writ of certiorari as improvidently granted) in which four members of the Court, including Chief Justice Rehnquist and Justice White who had earlier expressed reservations on the point, and Justice Powell, who did not participate in Tuttle, apparently agree that in appropriate cases, municipal liability might be based upon deficient training "policy".
. Our only published opinion to date dealing with a claim of municipal liability for a specific act of police brutality is Wellington, 717 F.2d 932. In that case we recognized the general viability of both theories, but concluded without full development of the theories that neither was supported by sufficient evidence on essential elements. See id. at 936 & n. 3 (custom); id. at 936-37 (policy). Cf. Milligan v. City of Newport News, 743 F.2d 227 (4th Cir.1984) (claim of deficient training of paramedical personnel dismissed on pleading; insufficient allegations of policy and causation).
. The Supreme Court’s grant of certiorari to review Kibbe v. City of Springfield, 777 F.2d 801 (1st Cir.1985), promised a definitive treatment of the deficient training theory that did not, however, eventuate. Following submission of the instant appeal to this court, the Court dismissed the Kibbe writ as improvidently granted. City of Springfield, Massachusetts v. Kibbe, — U.S. -, 107 S.Ct. 1114, 94 L.Ed.2d 293 (dismissed for inability to reach related question of whether more than negligence in training was required to establish municipal liability). This left municipal liability doctrine in the state of development that had been reached in Tuttle, 471 U.S. 808, 105 S.Ct. 2427, 85 L.Ed.2d 791, where the Court had purported to take only a "small” further step toward definitive development of the doctrine’s full contours, including the deficient training theory. Id. at 810, 105 S.Ct. at 2429.
. Whether the mere negligence of municipal policymakers in training their police is a sufficient degree of fault to impose municipal liability was left unresolved by the Supreme Court when it declined to decide the issue in Kibbe. See supra note 10. "Deliberate indifference" therefore remains the minimal standard in this circuit, on the authority of Wellington, 717 at 936 ("tacit authorization” or "deliberate indifference"). Though the question remains open in the Supreme Court, Justice O’Connor’s opinion dissenting from dismissal of the writ in Kibbe indicates that at least four members of the Court would impose a minimal standard of "reckless disregard for ... or deliberate indifference to [the rights of persons].’’ Kibbe, — U.S. at -, 107 S.Ct. at 1121; see also Fiacco v. City of Rensselaer, New York, 783 F.2d 319, 326 (2d Cir.1986) ("deliberate indifference’’); Languirand v. Hayden, 717 F.2d 220, 227 (5th Cir.1983) (same); Patzner v. Burkett, 779 F.2d 1363, 1367 (8th Cir.1985) (same).
. Another witness, Ammons, explained that this vulgarism was commonly used in the department to refer to a routine police practice of approaching any suspicious group activity by first indiscriminately manhandling its participants then attempting only later to sort out the innocent and the culpable, i.e., "take names.”
. The procedural implications of official capacity suits against municipal officials has created a more general confusion that has only recently been authoritatively resolved and that may understandably have been reflected in the party structuring and judicial handling of this lawsuit. See Brandon v. Holt, 469 U.S. 464, 471-73, 105 S.Ct. 873, 878-79, 83 L.Ed.2d 878 (1985) (implications and possible basis of confusion explained in light of evolving municipal liability doctrine since Monell)', see id. at 474-77, 105 S.Ct. at 879-81 (Rehnquist, J., dissenting) (majority’s analysis of procedural implications challenged).
. The City then pursues this analysis to claim that because the evidence clearly, established that City Manager Smith could not be charged with establishing or condoning the allegedly cul*1397pable "policy or custom” (as the jury later found) and the Civil Service Commission was not named as a defendant, there was no evidence that any official or agency with policy-making authority had acted culpably in the critical respect necessary to establish municipal liability. This ultimate argument fails for reasons explained in text.
. The district court explained to the jury that while it was to return verdicts respecting the culpability of each of the "representative” officials named as defendants, this would not result in individual liability on the part of any found culpable, but would impose liability "through them” on the City, and would also "specifically affix accountability."
It may be doubted that there is any technical basis for rendering the type of "merely declaratory” judgment against an official in his official capacity that the district court máy have had in mind. Any error in this procedure was surely harmless however. In a properly structured municipal liability action naming only the City as defendant, it would of course be perfectly appropriate under Fed.R.Civ.P. 49 to require a special verdict designed to establish which, if any, of various municipal officials might have been policymakers whose culpability in establishing policy or condoning custom could then be legally attributed to the City. In fact, this was the legal effect accorded by the district court to the jury's answers to the special verdict submission. As indicated in text, there was ample evidence supporting the jury’s determination that Chief of Police Dixon at least was a culpable policymaker whose culpability could fairly be attributed to the City "as its own.”
. It is not only without merit, but in one critical respect is positively disingenuous. In an obvious effort to bring the instructions here within those specifically disapproved in Tuttle, the City asserts in its brief that the district court instructed that the jury need only find that a proven policy or custom was “likely" or “might” lead to police misconduct such as that of McDaniel. Appellant’s Brief p. 18; cf. Tuttle, 471 U.S. at 824 n. 8, 105 S.Ct. at 2436 n. 8. The instructions nowhere use that term nor convey any comparable notion that so tenuous an “affirmative link” between policy or custom and specific violation would suffice in proof.
. In response to the defendants’ related objection that the hours recorded by Spell’s counsel reflected some duplication of effort, the court found that no evidence had been proffered which showed that duplicated time was unreasonable or unnecessary but nevertheless “employ[ed] a 5% across-the-board reduction in counsel’s requested ... hours in order to eliminate any unreasonable duplication that might have been overlooked, errors committed by this Court in denying defendants’ excessive hours objections ... and as a ‘practical means of trimming any fat from [plaintiffs] fee petition.’” 616 F.Supp. at 1094-95, partially quoting New York State Association for Retarded Children v. Carey, 711 F.2d 1136, 1146 (2d Cir.1983).
. This court has summarized the Johnson factors to include: (1) the time and labor expended; (2) the novelty and difficulty of the questions raised; (3) the skill required to properly perform the legal services rendered; (4) the attorney's opportunity costs in pressing the instant litigation; (5) the customary fee for like work; (6) the attorney’s expectations at the outset of the litigation; (7) the time limitations imposed by the client or circumstances; (8) the amount in controversy and the results obtained; (9) the experience, reputation and ability of the attorney; (10) the undesirability of the case within the legal community in which the suit arose; (11) the nature and length of the professional relationship between attorney and client; and (12) attorneys' fees awards in similar cases. Barber 577 F.2d at 226 n. 28.
.Although the court was unaware of any award of fees in a complex civil rights case within the district court in the last two years, Spell’s counsel provided four affidavits from members of the Raleigh and Fayetteville Bars indicating that in similar cases they, or their colleagues, would charge the same rates as those requested. 616 F.Supp. at 1105.
. The court expressly declined to make a further upward "quality adjustment,” 616 F.Supp. at 1110-11. This ruling is not challenged on appeal.
. Among the decisions relied upon was that of another district court of this circuit which was later affirmed by this court, see Vaughns v. Board of Education, 770 F.2d 1244, 1246 (4th Cir.1985). The district court’s allowance of a *1404multiplier was therefore well within contemplation of emerging circuit law at the time.
. The fee award at issue in Delaware Valley was made under § 304(d) of the Clean Air Act, 42 U.S.C. § 7604(d), but the Court’s decision plainly applies as well to awards under 42 U.S.C. § 1988. See Delaware Valley, — U.S. at -, 107 S.Ct. at 3085-86 (plurality opinion).
. Justice White, writing for four Justices, being "unconvinced that Congress intended the risk of losing a lawsuit to be an independent basis for increasing the amount of any otherwise reasonable fee,” Delaware Valley, — U.S. at -, 107 S.Ct. at 3086 (plurality opinion), concluded in a plurality opinion that "multipliers or other enhancement of a reasonable lodestar fee to compensate for assuming the risk of loss is impermissible under the usual fee-shifting statutes,” id. at-, 107 S.Ct. at 3088. Lacking a majority for this per se position, Justice White then concluded that even if contingency multipliers were permissible under the "typical fee-shifting statutes," they should only be allowed in “exceptional cases," and that on the facts of the particular case no enhancement was warranted. Id. at -, 107 S.Ct. at 3087-90. Four Justices, for whom Justice Blackmun wrote in dissent, disagreed, believing that fee enhancement for the risks of nonpayment was permissible in appropriate cases, and that remand for further fact findings would establish whether the case at issue was an appropriate one. Id. at -, 107 S.Ct. at 3091-3102 (Blackmun, J., dissenting). Justice O’Connor, writing separately, agreed with the dissenting Justices that Congress did not intend flatly to foreclose all consideration of contingency in setting reasonable fees under such fee-shifting statutes as 42 U.S.C. § 1988. But she agreed with the other four justices that a "risk multiplier” was not warranted in the circumstances of the case before the Court, id. at -, 107 S.Ct. 3089-92 thereby providing a majority for reversal. Justice O’Connor’s position on the circumstances under which a "contingency multiplier" may be appropriate is therefore the pivotal one.