SULLIVAN v. CITY OF SPRINGFIELD

LYNCH, Chief Judge.

Kevin Sullivan, Vincent Dudley, Mark Mehringer, Jason Sleeper, and Michael Trombley are current and former police officers who were hired by Springfield, Massachusetts ("the City") in 1997 and then laid off in 2003. Plaintiffs were all later recalled by the City. In 2005, they sued the City, asserting claims under 42 U.S.C. § 1983, inter alia.

Plaintiffs, who are white, claimed that "race-based determinations" made by the City during the hiring process in 1997 were not permitted under either a consent decree or the U.S. Constitution. They alleged these 1997 determinations harmed them by assigning them to lower seniority ranks than minority officers hired at the same time. Plaintiffs claimed that, if the City had not made these impermissible race-based determinations in 1997, they would have had higher seniority in 2003 and would not have been laid off or would otherwise have been recalled sooner.

On cross motions, the district court granted summary judgment to the City on two grounds. It found that plaintiffs had not shown facts establishing any race-based causal connection between defendant's actions and the plaintiffs' injuries. It also found the City's actions were within the scope of the City's ongoing obligations under a consent decree in a 1970s case known as Castro v. Beecher (Castro I), 334 F.Supp. 930 (D.Mass. 1971). See Sullivan v. City of Springfield, 555 F.Supp.2d 246, 256, 258 (D.Mass. 2008). We affirm the district court's decision on both grounds.

I.

A. The Castro Consent Decree

The history of the Castro consent decree sets the stage. In 1970, eight minority plaintiffs, who had applied unsuccessfully to become Boston police officers, brought suit against the Massachusetts Civil Service Division, an agency which has since become known as the Human Resources Division ("HRD"), the name we use. The suit alleged discriminatory hiring and recruiting practices, which violated the Fourteenth Amendment and had led to a disproportionately low number of minority police officers in Boston. See Castro I, 334 F.Supp. at 934-35; see also Deleo v. City of Boston, No. 03-12538, 2004 U.S. Dist. LEXIS 24034, at *4 (D.Mass. Nov. 23, 2004) ("At the time, the black population of [Boston] was approximately 16.3% of the total population, but about 3.6% of the Boston police force was black.").

The group was composed of six black applicants and two Hispanic applicants. For the purposes of the Castro decree and the surrounding litigation, the term "minority" referred only to black and Hispanic applicants, while "non-minority" referred to applicants of all other ethnicities; we use these terms accordingly in this opinion. See Quinn v. City of Boston, 325 F.3d 18, 24 n. 1 (1st Cir. 2003) (describing the use of the terms in parallel litigation involving firefighters).

For a more detailed history of the consent decree, see Deleo, 2004 U.S. Dist. LEXIS 24034, at *3-6, and Castro v. Beecher (Castro V), 522 F.Supp. 873, 874-76 (D.Mass. 1981).

The district court in Castro I rejected the plaintiffs' claims as to several of the eligibility requirements. See, e.g., Castro I, 334 F.Supp. at 940-41 (discussing the height and swim test requirements). It also declined to certify the plaintiffs' requested class. Id. at 947-48. It concluded, however, that the Massachusetts Civil Service Police Entrance Examination — the written examination administered by HRD to police applicants — discriminated against minorities who "did not share the prevailing white culture." Id. at 943. The court barred the use of the existing examination and set out guidelines for creating a non-discriminatory one. See id. at 944-45; see also Castro v. Beecher (Castro II), 459 F.2d 725, 729 (1st Cir. 1972).

The decision was appealed to this court, which held that class certification should have been granted. Castro II, 459 F.2d at 732. It agreed that the examination was discriminatory, but held that the district court's remedy had been too narrow, and remanded. This court stated that, "[i]n our view, if relief in the near future is to be more than token, further provision is necessary," and that the relief could include the creation of a priority pool for minority applicants who passed a non-discriminatory examination and who could be hired according to a preferential ratio. Id. at 737. While recognizing that "any such effort is bound to be a crude one and must be pursued with sensitivity," this court stated that "preferential status for the priority pool will yield a significant increment of [minority] police officers in the near term." id at 736-37.

Following the remand, and in the context that use of the extant examination was illegal, the parties entered into what has now come to be known as the Castro consent decree. The decree was approved by the district court in 1973. Castro v. Beecher (Castro III), 365 F.Supp. 655, 660 (D.Mass. 1973). The court stated that the decree was intended to "counteract the unconscious lopsidedness of the recruitment of the past" by "giv[ing] a . . . priority to [minority candidates] who have shown themselves qualified." Id. at 659. In entering a consent decree, the defendant state authorities may well have agreed to relief beyond what the Constitution would have provided as a remedy. See United States v. Charles George Trucking, Inc., 34 F.3d 1081, 1091 (1st Cir. 1994). This is a point we need not decide.

The consent decree was subsequently revisited and modified. See Castro V, 522 F.Supp. at 875; Castro v. Beecher (Castro IV), 386 F.Supp. 1281, 1285-86 (D.Mass. 1975). The amended remedy was explicitly modeled on the consent decree adopted in a parallel case involving Massachusetts firefighters. See Castro IV, 386 F.Supp. at 1286 (citing Boston Chapter, NAACP, Inc. v. Beecher, 371 F.Supp. 507 (D.Mass.), aff'd, 504 F.2d 1017 (1st Cir. 1974)); see also Castro V, 522 F.Supp. at 875.

The Castro consent decree required HRD to prepare certification lists by creating two groups. The first, "Group A," would "consist of all Black and Spanishsurnamed applicants who pass a future police entrance examination and are otherwise qualified for appointment on the basis of existing requirements." The second, "Group B," would "consist of all other persons who pass a future police entrance examination and are otherwise qualified for appointment on the basis of existing requirements." When an appointing authority, such as the City, sought to hire police officers, it would send a request to HRD, which would send the appointing authority a certification list ordered "on the basis of one candidate from Group A for every candidate certified from Group B." In making its hiring decisions, if the appointing authority chose to reject a candidate in favor of another candidate who appeared lower on the HRD list, HRD would not approve the appointment unless the appointing authority "furnished [HRD] with a written statement of [its] reasons" for doing so; HRD would then provide a "written statement of those reasons to . . . the candidate upon written request."

The Castro decree was to remain in effect for a given city until that city's police department "achieves a complement of minorities commensurate with the percentage of minorities within the community." The decree accordingly continues to apply to Springfield. The plaintiffs agree that is so. Boston is no longer under the decree. See Deleo, 2004 U.S. Dist. LEXIS 24034, at *26 (finding that Boston had reached this goal for minority representation).

B. Facts Underlying the Present Case

The facts in the present case are undisputed, and it is undisputed that there are gaps in the evidence about plaintiffs' hirings and layoffs. This may in part be attributable to the fact that plaintiffs did not bring suit until 2005, some eight years after they were hired.

Because we affirm on other grounds, we do not consider defendant's argument that plaintiffs' action was barred by the statute of limitations.

In 1996, Springfield decided to hire sixty new police officers. Its police commission chairman requested a list of eligible candidates from HRD. Eligibility meant something more limited however; in the intervening years, HRD had delegated to the local appointing authorities most of the responsibility for determining the eligibility of candidates for eligibility criteria other than the examination.

On January 13, 1997, in response to the City's request, HRD generated a list, Certification No. 961295, of 374 candidates who had passed a written examination. Under the Castro decree, the list placed the minority candidate who had scored highest on the test first, followed by the non-minority candidate who had scored highest on the test, followed by the second highest scoring minority candidate and the second highest scoring non-minority candidate, and so on, in alternating order.

The HRD list did not include applicants' scores. As a result, there is no direct evidence that minority individuals with lower exam scores had higher positions on the list than higher scoring non-minorities.

After the City received the HRD examination certification list, it asked those candidates who were still interested in being considered for the job to come in and sign the certification list. The record does not establish which candidates signed the list, and which were eliminated from consideration for failure to do so. Although an appointing authority such as the City would generally have sent the original, signed certification list back to HRD, the relevant certification list — or other evidence as to which candidates signed and which did not — is not in the record.

HRD retained oversight of the process used by local appointing authorities to ensure compliance with its rules. The City applied a series of hiring prerequisites to the remaining candidates to decide which candidates would receive offers and be able to go through the police academy. For each candidate, the City ran a background check, a residency check and an interview; if the candidate was not eliminated at this point, he or she would be given a conditional offer of employment, after which he or she would be further subject to a medical examination, a psychological examination, and a physical abilities test.

Candidates were eliminated if they did not have or could not obtain a Massachusetts driver's license, or had felony convictions, or lied about being Springfield residents.

Captain William Cochrane, who from 1996 to 2001 was the commanding officer of the Academy for the Springfield Police Department, was responsible for this process. Cochrane and his staff evaluated and eliminated candidates on the basis of the qualifications listed above. Once they had a list with enough qualified candidates to meet the City's hiring goals, Cochrane passed this list to the chiefs office, which then filled out an Authorization of Employment Form 14 ("Form 14"). Form 14, which was sent to the state HRD on May 5, 1997, contained the names of officers who had been selected for appointment. In addition to candidates who were on the HRD certification list, Form 14 included the names of eight former police cadets as well as two other officers whose employment applications were processed differently.

Cochrane testified at deposition that in carrying out this process, he and his staff attempted to comply with the requirements of the Castro decree. To this end, he had spoken with HRD "many, many times to try and clarify how we were to proceed." Cochrane testified that, in order to comply with what he understood as the requirements of the Castro decree, after he received the original list from HRD, he and his staff worked off of two separate lists — one with the minority candidates in ranked order and the other with the non-minority candidates in ranked order. If a minority candidate was eliminated from consideration under the qualification procedures, Cochrane and his staff would proceed to considering the next minority candidate. Likewise, if a non-minority candidate was eliminated, Cochrane and his staff would proceed to considering the next non-minority candidate. Cochrane testified that his memory was that, at the end of the process, these two lists were combined in alternating order — one minority candidate followed by one non-minority candidate. That racially reordered list was then given to the chiefs office, which filled out Form 14 accordingly. That Form 14 was placed into evidence. The Form 14 shows that the City did not hire on a one-for-one minority-to-non-minority basis. Rather, in the spring of 1997, the City hired about twice the number of white officers as minority officers, excluding police cadets. If cadets are counted, the City hired more than twice the number of white officers as minority officers.

Contrary to Cochrane's memory, Form 14 was not in fact organized on a one-to-one minority-to-non-minority basis. The district court justifiably concluded that Cochrane's memory was simply incorrect. Sullivan, 555 F.Supp.2d at 253. Even discounting the eight cadets and the two other candidates who were treated separately, Form 14 contained twenty-one non-minority candidates and twelve minority candidates. The minority candidates are not in alternating order. Within each group, the candidates were in the same rank order as they were on the original HRD list. But some non-minority candidates, including the plaintiffs, appeared lower on Form 14, relative to some minority candidates, than they did on the original HRD list. As the district court noted, there is no evidence as to when (if ever) the City created one or more reordered one-for-one lists. And there is no evidence that the ordering on the list was based on race.

The City assigned badge numbers according to the order on Form 14. It also simultaneously assigned seniority according to officers' badge numbers; the greater the officer's seniority, the lower the badge number. Thus, those who appeared lower on Form 14 were assigned less seniority than those who appeared higher on Form 14. Plaintiffs' seniority dates were assigned on this basis.

Almost six years later, in early 2003, the City laid off approximately seventy-five police officers. It is undisputed that the layoff order was determined purely by seniority — that is, by badge number, and not by race. The City laid off the five plaintiffs over a two-week period beginning March 13, 2003. Later that year, and through the end of 2004, the City obtained more funding through grants and retirements, and was able to recall all of the laid-off police officers who wanted to return. As a result, plaintiffs were all re-called. It is undisputed that the recall order was also based on seniority, not race. Because both the layoffs and the recalls were conducted on the basis of seniority, officers with less seniority were laid off sooner and recalled later than those with more seniority.

On January 11, 2005, plaintiffs brought their § 1983 suit, claiming that they were assigned higher badge numbers in 1997, and thus were given less seniority, on the basis of impermissible race-based determinations. That is, they argued that if the City had not racially reordered the HRD list as it was going through the various qualification procedures, they would have been hired with greater seniority, and thus would have been either not laid off at all in 2003, or would have been recalled more quickly than they were. As the district court viewed their claim, the plaintiffs were arguing that the "failure to assign seniority based on the original, alternating HRD certification, . . . instead [of] the Form 14, violated both the Castro decree and the Equal Protection Clause." Sullivan, 555 F.Supp.2d at 255.

Plaintiffs alleged that as a result of the City's actions, they lost, among other things, back pay, overtime, details, and health insurance contributions. They sought compensatory damages, back pay, front pay, and compensation for lost overtime and details. Plaintiffs' claims of harm necessarily rest on their contention that the original assignment of seniority was discriminatory. They do not challenge the validity or constitutionality of the City's continuing obligations under the Castro decree, but argue that the City's actions were not within the scope of the decree.

On November 3 and December 19, 2005, the parties filed cross motions for summary judgment. On January 3, 2007, the district court denied both motions without prejudice, citing certain gaps in the record, but invited the parties to file substitute motions that resolved these gaps. Plaintiffs filed a new motion for summary judgment on May 14, 2007, which attempted to clarify the record. The City filed a cross motion for summary judgment on September 5, 2007. On March 28, 2008, the district court denied plaintiffs' motion and granted defendant's motion; the court issued a second memorandum on May 23, 2008 reiterating its conclusion and explaining its reasoning more fully. Sullivan, 555 F.Supp.2d at 246. The court noted that there was no material dispute of fact. Id. at 247. It found that plaintiffs had not established the causal connection between defendant's actions and their injuries because there appeared to be no racial reordering of Form 14. See id. at 253-54. Furthermore, the court held that even if this causal connection could be established, the City's actions were within the scope of the Castro decree. See id. at 258.

Plaintiffs timely appealed.

II.

We review de novo the grant of summary judgment. Rodriguez v. Am. Int'l Ins. Co. of P.R., 402 F.3d 45, 46 (1st Cir. 2005). All reasonable inferences must be drawn in favor of the non-moving party; still, we ignore "conclusory allegations, improbable inferences, and unsupported speculation." Prescott v. Higgins, 538 F.3d 32, 39 (1st Cir. 2008) (quoting Medina-Munoz v. R.J. Reynolds Tobacco Co., 896 F.2d 5, 8 (1st Cir. 1990)) (internal quotation marks omitted). Summary judgment is appropriate where "there is no genuine issue as to any material fact" and "the movant is entitled to judgment as a matter of law." Fed.R.Civ.P. 56(c); see also New Fed Mortgage Corp. v. Nat'l Union Fire Ins. Co. of Pittsburgh, PA, 543 F.3d 7, 11 (1st Cir. 2008).

Plaintiffs do not challenge the application of the Castro decree by the City. Cf. Mackin v. City of Boston, 969 F.2d 1273, 1275 (1st Cir. 1992). They concede that to the extent defendant's actions were within the scope of the decree, they have no claim for relief. Plaintiffs' argument is that defendant made race-based employment decisions that went beyond the dictates of the Castro decree. These actions, they urge, must be subject to strict scrutiny. See Parents Involved in Cmty. Schs. v. Seattle Sch. Dist. No. 1, 551 U.S. 701, 127 S.Ct. 2738, 2751, 168 L.Ed.2d 508 (2007) ("It is well established that when the government distributes burdens or benefits on the basis of individual racial classifications, that action is reviewed under strict scrutiny."). Plaintiffs argue that defendant's actions cannot withstand strict scrutiny.

Plaintiffs' claims focus on the actions of Cochrane. In his deposition, Cochrane stated that he and his staff reordered the list sent to the City by HRD in 1997 in order to maintain the one-to-one minority-to-non-minority order as candidates were eliminated from the list. If a minority candidate was eliminated, he was replaced by the next highest minority candidate; if a non-minority candidate was eliminated, he was replaced by the next highest non-minority candidate. Plaintiffs claim that because of this race-based reordering, they were hired with higher badge numbers — that is, with less seniority — than they would have had if hiring proceeded purely according to the order of the HRD list. And although Cochrane claimed the reordering was done to comply with the Castro decree, plaintiffs argue the decree did not allow any race-based ordering beyond that already done by HRD.

A. Causation

Plaintiffs' opening argument on appeal is that they are entitled to strict scrutiny of any governmental decision based on race. See Parents Involved, 127 S.Ct. at 2751; Johnson v. California, 543 U.S. 499, 505-06, 125 S.Ct. 1141, 160 L.Ed.2d 949 (2005). However, they have failed to produce evidence that the governmental decisions here were based on race.

In order to have a valid claim under § 1983, plaintiffs must show that defendant's actions were the cause in fact of the alleged constitutional deprivation. Gagliardi v. Sullivan, 513 F.3d 301, 306 (1st Cir. 2008) (citing Rodriguez-Cirilo v. Garcia, 115 F.3d 50, 52 (1st Cir. 1997)); see also 1 S. Nahmod, Civil Rights and Civil Liberties Litigation § 3:110 (4th ed. 2008). It is not enough for plaintiffs to show Cochrane may have used an impermissible racial classification; there must be a causal link between this and the adverse employment action — that is, plaintiffs' being given their particular seniority ranking on hiring. See Gagliardi 513 F.3d at 306-08; Back v. Hastings on Hudson Union Free Sch. Dist., 365 F.3d 107, 125 (2d Cir. 2004) ("[On an alleged Equal Protection Clause violation], the plaintiff must show more than invidious intent. She must also `demonstrate that the causal connection between the defendant's action and the plaintiffs injury is sufficiently direct.'" (quoting Gierlinger v. Gleason, 160 F.3d 858, 872 (2d Cir. 1998))). To survive summary judgment, plaintiffs were required to offer sufficient evidence to allow a reasonable fact finder to conclude defendant's actions were based on race and this caused their injuries. See Soto v. Flores, 103 F.3d 1056, 1065-66 (1st Cir. 1997); see also Rodriguez-Cirilo, 115 F.3d at 52-53; Back, 365 F.3d at 125-26. Plaintiffs have failed to do so.

To the extent plaintiffs continue to argue that the City was obliged not to alter the HRD list or that any alterations to it were race-based, the argument is simply wrong. The HRD list was based only on examination results; all other eligibility criteria were left to the City to determine and on that basis the list was altered, introducing a large number of variables other than race.

It is undisputed that seniority was assigned purely on the basis of Form 14, which the City sent to HRD on May 5, 1997. It is also undisputed that plaintiffs' names were lower on Form 14, relative to certain minority candidates, than they were on the original list HRD sent to the City. But the plaintiffs have not shown this was based on race. Plaintiffs introduced evidence — Cochrane's deposition — that the City at some point reordered the HRD list on the basis of race as it was going through the qualification process. There is no evidence in the record, however, beyond plaintiffs' speculation, from which a fact finder could reasonably determine what effect Cochrane's reordering of the list the City received from HRD had on the list that was ultimately sent back to HRD, Form 14. There is a break in the causal chain between the original HRD list, Cochrane's reordering, and Form 14.

Plaintiffs rely on Cochrane's testimony that he and his staff split the original HRD list into two lists, one for minorities and one for non-minorities; that they worked off these separate lists during the qualification process; and that they then created the final list that was the basis for Form 14, at the end of the process, by recombining the two lists, dovetailing minority and non-minority candidates.

But the record evidence does not stop with Cochrane's account. The actual order of persons on the Form 14 sent by the City to HRD is in evidence and it does not correspond to Cochrane's description. Discounting the eight cadet candidates and candidates Jose Diaz and Brian Elliot, who were all handled under separate procedures, the names on Form 14 were not organized in one-to-one minority-to-non-minority order; nor are there equal numbers of minority and non-minority candidates. Thus, it cannot be that Form 14 was created simply by combining the two race-segregated lists.

Even assuming a race-based, one-to-one interim list was created at some point in the process, plaintiffs have presented no evidence they were lower on this list than they were on the HRD list, or that this was ultimately the reason why plaintiffs were lower on the final Form 14 list. Inferences must be reasonable ones and it is plaintiffs' burden to present enough evidence for a trier of fact to infer that there was a causal connection as to both lists. Speculative hypotheses are not enough.

We agree with the district court that plaintiffs' speculation that there may be a causal chain, without supporting evidence, is insufficient to survive summary judgment. See Gagliardi, 513 F.3d at 307 (upholding dismissal under Fed.R.Civ.P. 12(b)(6) of a § 1983 suit where plaintiffs theory of causation was "conclusory and . . . not substantiated by reasonable inference from the well-pleaded facts"); Rodriguez-Cirilo, 115 F.3d at 53 (upholding grant of summary judgment in § 1983 suit in which plaintiffs "offered no competent evidence that could have supported a finding" of but-for causation).

Plaintiffs try a different argument: that summary judgment was premature because there are documents that could be brought before the court that could prove the causal connection. There are several reasons to reject the argument.

Plaintiffs are correct that such documents at one point existed, but they have not produced them. For example, at some point after receiving the original HRD list, but before filing the final Form 14, the City would have filed with HRD a Form 1611; this form would have indicated which candidates were removed from the list for failure to appear and sign the initial certification or eliminated for failure to meet one of the other requirements for employment. The City would have also returned to HRD the original certification list, signed by those applicants willing to accept certification. Beyond this, there is the final list that Cochrane would have given the chief's office, and any interim lists that Cochrane and his staff may have created.

First, not only did the plaintiffs not argue to the district court that summary judgment was premature, they affirmatively requested that the court resolve the case on the existing evidence; plaintiffs filed their summary judgment motion before the City filed its cross motion for summary judgment. Plaintiffs thought the statements in Cochrane's deposition were sufficient not only to defeat defendant's summary judgment motion but to establish their case as a matter of law. The district court disagreed, and "plaintiff[s'] disappointment with that ruling is understandable," but there is no reason for us now to "turn back the clock and give [them] a further opportunity to reconfigure the record." Vargas-Ruiz v. Golden Arch Dev., Inc., 368 F.3d 1, 5 (1st Cir. 2004); see also Meehan v. Town of Plymouth, 167 F.3d 85, 92 n. 7 (1st Cir. 1999).

Second, the fact that such documents at one point existed and would likely have been in the possession of defendant or HRD did not relieve plaintiffs of their burden to produce evidence necessary to support a jury verdict, see Singh v. Blue Cross/Blue Shield of Mass., Inc., 308 F.3d 25, 32-33 (1st Cir. 2002). Plaintiffs never filed a Rule 56(f) affidavit or otherwise represented that consideration of summary judgment was premature. See Fed.R.Civ.P. 56(f) (providing that, "[i]f a party opposing [a summary judgment] motion shows by affidavit that, for specified reasons, it cannot present facts essential to justify its opposition, the court may . . . order a continuance to enable . . . other discovery to be undertaken"). Plaintiffs had a full opportunity to conduct discovery. See Vargas-Ruiz, 368 F.3d at 3 (stating that when "a party professes an inability to respond to a summary judgment motion because of incomplete discovery, his recourse is by way of Fed.R.Civ.P. 56(f)"). Plaintiffs cannot claim to be surprised that the procedures by which the City handled HRD certification list No. 961295 and the effect these procedures had on plaintiffs were at issue; indeed, these issues were central to the case from the start.

Plaintiffs counter that the necessity of such evidence only became apparent when the district court, in ruling on the summary judgment motions, made reference to two documents not contained in the record — an unpublished stipulation of facts between the parties in the original Castro v. Beecher case and a Certification Handbook prepared by HRD. The court's use of these documents will be discussed further below, but plaintiffs' argument fails in this context. The documents used by the district court were unrelated to the question of whether Cochrane's actions could be established as the cause in fact of plaintiffs' injuries.

B. Scope of the Consent Decree

The district court held that the City's actions were part of its obligations under the Castro consent decree. Plaintiffs argue that defendant's actions went beyond the scope of the Castro decree, which they say should be narrowly read. They argue the requirements of the Castro decree were fulfilled when, following the written examination, HRD created a list of candidates that alternated minority and non-minority candidates; any race-based reordering by the City after it received the list from HRD was impermissible.

Plaintiffs offer two arguments in support of their interpretation of the consent decree. First, because the consent decree was intended to remedy the effects of a discriminatory written examination, it contemplated race-based reordering only after that stage of the process, and not also after candidates were eliminated in later stages, such as the medical or psychological examinations. Second, even if the consent decree did allow further reordering, such reordering could be done only by HRD, and not by the City. We disagree on both counts, as the district court did, see Sullivan, 555 F.Supp.2d at 256-58.

The language of the consent decree requires the creation of two separate lists: one of minority applicants "who pass a . . . police entrance examination and are otherwise qualified for appointment on the basis of existing requirements"; and another of non-minority candidates "who pass a . . . police entrance examination and are otherwise qualified for appointment on the basis of existing requirements." (Emphases added.) It then specifies that "candidates shall be certified on the basis of one candidate from [the minority list] for every candidate certified from [the non-minority list]."

The language of the decree, evidence concerning past practices under the decree, and common sense regarding the intended operation of the decree are relevant considerations in interpreting the scope of the Castro consent decree. See Mackin, 969 F.2d at 1276. All these considerations support the conclusion that the City's actions were well within the scope of the Castro decree.

Looking at the language of the consent decree, the most natural reading does not support plaintiffs' contention that race-based reordering under the decree can occur, only once, immediately following the written examination. The parties to the Castro decree agreed to a remedial approach that is not strictly limited to addressing the written examination problem. The language of the Castro decree clearly allows for race-based reordering to occur not merely after the written examination, but also after candidates have been evaluated and eliminated on the basis of other qualifications.

It is true that the litigation that led to the creation of the Castro decree was focused on the discriminatory effects of the written examination. See Castro III, 365 F.Supp. at 655-56; see also Castro II, 459 F.2d at 735-38. However, because consent decrees are "animated not only by the parties' legal claims but also by the parties' consent," it is not impermissible for a consent decree to extend beyond the "possible bounds of a decision issued directly by the trial court." Charles George Trucking, 34 F.3d at 1091.

The language of the Castro decree also does not say that HRD could not delegate its authority to carry out such reordering to the City. Indeed, the decree explicitly provides for cooperation between HRD and the coordinating authority in carrying out the decree. Once HRD has created the original certification list, an appointing authority may still "reject[] the appointment of [a] candidate higher on the list" in favor of a candidate lower on the list if it "furnishes [HRD] with a written statement of [its] reasons" for doing so. Such language is not consistent with plaintiffs' claim that the drafters of the Castro decree intended to strictly limit HRD's authority to share its responsibilities under the decree with appointing authorities.

Past practices do not support either plaintiffs' argument as to the order in which other qualifications could be determined or their non-delegation argument. Early on, it was recognized that HRD could take other qualifications, beyond the written exam, into account before creating the alternating list. For example, in approving the original consent decree, the district court in Castro III stated that the list of minority applicants who passed the 1972 written examination would be "exclusively composed of persons who have a high school diploma" and who "measure up prima facie to reasonable requirements for the job of police patrolman." Castro III, 365 F.Supp. at 658. Two years later, the district court in Castro IV recognized that statutory residence preferences were preserved under the Castro decree, but that this preference "and like rankings," Castro IV, 386 F.Supp. at 1285 (quoting Castro III, 365 F.Supp. at 658), are to be taken into account only "within each of the four groups established by the decree," id.

Further, it was recognized in the litigation leading up to the creation of the Castro decree that HRD and the appointing authorities would have to work together in order to effectuate the decree. See Castro II, 459 F.2d at 737 ("[T]he remedial approach . . . [of the decree] cannot be fully effective without the cooperation of the appointing authorities in the various police forces."). This high degree of cooperation is apparent in Cochrane's statement that HRD had spoken with him "many, many times" in order to "clarify how we were to proceed" under the Castro decree.

Plaintiffs argue that we should not look to past practices in this case because the City was not a party to the original Castro litigation, so its actions in following the decree do not evidence the intent of the decree's framers. Regardless of whether the City was a party to the original litigation, the past practices we have examined are not those of the City but those of HRD — in taking other qualifications into account prior to creating the alternating list and in sharing its responsibilities with appointing authorities. HRD was a party to the original litigation. Nor is evidence of past practices made irrelevant by the fact that practices have changed over time.

Finally, the logic inherent in the decree undercuts plaintiffs' argument. Although the Castro decree was based on a finding that the written entrance examination was discriminatory, see Castro II, 459 F.2d at 735-36, under our circuit precedent, fixing the problem of discriminatory examinations was "not an end in [itself] but merely a means toward achieving the decree's actual objective: rough parity," Mackin, 969 F.2d at 1277. Our decision in Castro II rejected the argument that the decree was intended only to address problems related to the written examination, as plaintiffs claim, noting that "if relief in the near future is to be more than token, further provision is necessary." Castro II, 459 F.2d at 737; see also Mackin, 969 F.2d at 1277 n. 3.

Since the Castro decree allowed HRD to reorder the certification list based on race after taking into account qualifications beyond the written examination, moreover, there is no reason to think that it would bar HRD from delegating the same task to an appointing authority such as the City. See Mackin, 969 F.2d at 1276 ("We think it is farfetched to assume that the district court or the parties intended the decree to work in so quirky a fashion.").

We need not consider the City's argument that liability cannot attach because there is no evidence that Cochrane was acting out of any motivation other than his desire to adhere to his understanding of the legal requirements of the Castro decree.

III.

Plaintiffs also argue that the district court impermissibly went beyond the record in issuing its summary judgment ruling. Plaintiffs' claim focuses on the court's use in its May 23, 2008 order of two documents, neither of which had been presented by the parties. First, the court referenced an unpublished stipulation of facts between the parties in the original Castro v. Beecher case. Second, the court referenced a Certification Handbook prepared by HRD to give guidance to appointing authorities on hiring under the Castro decree; the Handbook was apparently prepared after the hirings at issue in this case took place. Plaintiffs argue that the court could not base its summary judgment conclusion on this new evidence, which plaintiffs did not have the opportunity to challenge or overcome.

Even if the court should have taken judicial notice of these documents before it considered the summary judgment motions, see generally 10A Wright, Miller Kane, Federal Practice and Procedure § 2723, at 391-97 (3d ed. 1998), there was no prejudice to the plaintiffs. There was no reversible error because "[t]he critical facts upon which the court relied were either squarely presented or plausibly inferable" from the undisputed facts. Forcier v. Metro. Life Ins. Co., 469 F.3d 178, 186 (1st Cir. 2006) (emphasis added). Neither document was an essential ingredient of the analysis, which would have reached the same result without them.

IV.

Entry of judgment for the City is affirmed.