Fisher v. Koehler, 692 F. Supp. 1519 (S.D.N.Y. 1988)

US District Court for the Southern District of New York - 692 F. Supp. 1519 (S.D.N.Y. 1988)
August 9, 1988

692 F. Supp. 1519 (1988)

Charles FISHER, Edward Holohan, John Harris, Ledgure Davis, Lionel Lawson, Victor Ellis, John Thompson, Milton Gadson, individually and on behalf of all other persons similarly situated, Plaintiffs,
v.
Richard KOEHLER, Commissioner of the Department of Correction of the City of New York; James T. Garvey, Supervising Warden of C-76; Bruce Sullivan, Warden of C-76; Sheila Vaughn, Deputy Warden for Administration at C-76; David C. White, Deputy Warden for Security at C-76; Joseph Colon, Deputy Warden for Programs at C-76; Steven Thomas, Assistant Commissioner for Health, Education and Social Services for the Department of Correction; Steven C. Joseph, Commissioner of the Department of Health; Allan Goldberg, Director of Prison Health Services for the Department of Health; Mel Kaye, Ph.D., Director of Mental Health services for Prison Health Services; Dr. Murray Rosenthal, Director of Dentistry of the Department of Health; Edward Koch, Mayor of the City of New York, Defendants.

No. 83 Civ. 2128 (MEL).

United States District Court, S.D. New York.

July 13, 1988.

As Amended August 9, 1988.

*1520 *1521 Philip L. Weinstein, John Boston, Claudette R. Spencer, Dale A. Wilker, The Legal Aid Society, Prisoners' Rights Project, New York City, for plaintiffs.

Peter L. Zimroth, Corp. Counsel of the City of New York, New York City (Diane Morgenroth, Steven H. Mosenson, Hway-Ling Hsu, Asst. Corp. Counsels, of counsel), for defendants.

LASKER, District Judge.

In this action, plaintiffs, who are a class[1] of inmates at the New York City Correctional Institution for Men ("CIFM"),[2] have challenged the conditions of confinement at CIFM on many grounds, of which rampant violence, both between inmates and between staff and inmates, is one of the most significant and distressing. After a long trial on a motion for a permanent injunction on the separate issue of violence at CIFM and its causes, it is determined that plaintiffs have established that violence at CIFM, both inmate-inmate and staff-inmate, has reached proportions that violate the Eighth Amendment.[3] Systematic deficiencies in the operation of CIFM, most significantly, overcrowding, overreliance on open dormitory housing, lack of adequate classification, inadequate staffing and supervision, and inadequate systems for controlling, investigating and disciplining misuse of force, have led to a world where inmates suffer physical abuse, both by other inmates and by staff, in a chillingly routine and random fashion. Although defendants have implemented many positive changes at CIFM during the course of and after the trial, the entrenched and recurring nature of violence at CIFM requires the issuance of an injunction, albeit one which will take into account defendants' recent efforts to improve the situation.

On October 3, 1986, plaintiffs filed a motion for a preliminary injunction "requiring defendants to take appropriate action to end the pattern of violence among inmates and between inmates and staff at" CIFM,[4] one of the most pressing issues raised in the complaint. The hearing on the preliminary injunction began on October 23, 1986; the hearing on the application for a preliminary injunction was later consolidated with the trial on the merits on the separate issue of violence, under Fed.R. Civ.P. 65(a) (2). Testimony was heard at intervals through June 9, 1987, and the record was substantially completed on June 25, 1987. Post-trial memoranda, totalling almost 950 pages, were submitted by October 26, 1987.

Plaintiffs' witnesses consisted of fourteen CIFM inmates and former inmates and four expert witnesses. Toni V. Bair, regional administrator of the Central Region of the Virginia Department of Corrections, and Colonel James Shoultz, who has most recently been Director of Corrections in the Seminole County, Florida jail system, testified as experts on correctional administration.[5] Vincent M. Nathan, an attorney *1522 who has served as a court-appointed master in numerous cases involving the misuse of force in prisons, testified as an expert concerning use of force issues, including the investigation, monitoring and discipline of use of force, and remedies for use of force in correctional systems. T. 4233.[6] Professor Verne C. Cox, chair of the Department of Psychology of the University of Texas at Arlington, testified as an expert in psychology and the effects of crowding.[7]

Defendants' witnesses included Richard J. Koehler, Commissioner of the New York City Department of Correction ("Department of Correction"); James T. Garvey, who was then Warden of CIFM;[8] Judy Schultz, who was then Inspector General of the Department of Correction; Francesca Digirolamo, the civilian Director of Classification; *1523 a former director from the Training Academy; five CIFM captains and ten CIFM correction officers. Defendants called two expert witnesses. James W. Painter, who is Chief of the Custody Division of the Los Angeles County Sheriff's Department, testified as an expert on corrections.[9] Dr. Gerald G. Gaes, a Senior Research Analyst at the United States Bureau of Prisons, testified as an expert on the effects of prison crowding.[10]

Both parties also submitted extensive documentary evidence at trial. In addition, following the testimony of the expert witnesses, the parties entered into two testimonial stipulations. Defendants' Exhibit ("DX") XXX describes further testimony that certain witnesses for the defendants would have presented if they had been called, including efforts at improvement in various areas at CIFM. DX DDD describes new procedures and efforts at improvement in the Department of Correction's classification system about which Francesca Digirolamo, defendants' Director of Classification, would have testified if she had been recalled as a witness. Finally, with the submission of post-trial briefs, the parties also submitted further affidavits, including a supplemental affidavit from Commissioner Koehler describing defendants' efforts to improve CIFM since the close of the trial.

The court, accompanied by counsel for the parties and various CIFM officials, toured CIFM on January 8, 1987, visiting many of the areas relevant to the case.

 
I. The New York City Correctional Institution for Men

CIFM is a medium security facility operated by the New York City Department of Correction. It is the principal facility in which New York City's sentenced male inmates are incarcerated. The city-sentenced inmates at CIFM are serving terms of one year or less for violations, misdemeanors, or low-degree felonies.[11] Of the sentenced men, roughly eighty-two percent are adults aged twenty-one and above, and approximately eighteen percent are adolescents aged sixteen to twenty. CIFM also houses technical state parole violators who have been remanded pending a final revocation hearing; these violators constitute roughly seventeen percent of the population.

CIFM's official capacity is 2083. As is discussed in greater detail in Section IV of this opinion, its population has reached peaks as high as 2800; the average population at CIFM at the time of the trial was about 2500-2600 inmates. Commissioner *1524 Koehler testified that the average length of stay for sentenced inmates at CIFM is fifty days, T. 3434, and the average length of stay for parole violators is fifty-four days, T. 3435, but many of the inmates who testified at trial had been incarcerated for much longer periods, several for as long as eight months, and one for over a year. (Consecutive sentences may result in an inmate's incarceration at CIFM for more than a year.) Other inmates are at CIFM for far briefer periods than fifty days.

CIFM's main building was built in 1964, with one set of dormitories added in 1970. Three prefabricated "modules" were attached to the jail in 1984-1985. General population inmates at CIFM, as well as most protective custody inmates, are housed in large unpartitioned dormitories. PX 247 (admissions 24 and 31); T. 1798 (Captain DeCicco). In addition, there are four cell areas, with a total of 136 cells, used for adult administrative segregation, adolescent administrative segregation, mental observation, and punitive segregation. There are also dormitories designated for drug detoxification, mental observation, and infirmary use.

On the North Side of CIFM there are two floors of twin cell corridors in the shape of a chevron; there are thirty-four cells on each corridor, for a total of 136 cells, as stated above. Each cell has a window, a toilet and a sink; most of the cell areas have a dayroom at the front. There are also four dormitories on the North side: two per floor. The South Side of CIFM and the Annex each contain three floors with four dormitories per floor. These dormitories are large rectangular areas with beds, most of them double bunks, and with a dayroom, bathroom and officer's station near the entrance. The three modular units each contain two separate prefabricated housing units, which share a single large officer's station near the entrance.[12]

CIFM also contains dining areas for inmates and staff, a chapel, a gymnasium, a clinic, a law library and other program and administrative areas.

 
II. Inmate-Inmate Violence at CIFM: Scope of the Problem

Plaintiffs have established that inmate-inmate violence pervades CIFM. As described below, the evidence at trial demonstrated an alarmingly high level of reported incidents of inmate-inmate violence, which has spiralled upward from almost nine hundred incidents in 1982 to over thirteen hundred incidents in 1986. Although about twenty-five percent of these incidents were fights in which there was no documentation of injury, plaintiffs have documented numerous examples of stabbings, burnings, sexual assaults and other serious and gruesome injuries. Many of these injuries were inflicted by inmates who had acquired lengthy histories of violence against other inmates but who were still housed and allowed to roam at large in open dormitories. Violent injuries took place not only in housing dormitories, but in cell areas, in the halls, on staircases and even in protective custody areas, where inmates who fear for their safety seek refuge.

 
A. Eyewitness Testimony

Twelve inmate witnesses testified about inmate-inmate violence at CIFM. Each inmate testified to a number of separate incidents of such violence, in which the inmate had either been the victim or an eyewitness. Some of the incidents described bordered on the trivial, and a few were incredible altogether or in part. However, taken as a whole, the inmate testimony, which was supported in many instances by documentary evidence, credibly describes CIFM as an institution where inmate-inmate assaults and fear of assault are rampant. Defendants, for the most part, did not challenge the essence of the inmates' accounts of these incidents, except to dispute in some instances the exact circumstances of the injury, such as whether a weapon *1525 was involved. For the purpose of concluding whether plaintiffs have established a pattern of pervasive inmate-inmate violence, the fact that each incident occurred, the seriousness of any injury, and the place where the incident occurred are the most important considerations, and thus I have not attempted to resolve all inconsistencies in testimony.

Finally, it should be noted that not all the inmate testimony cast CIFM correction officers in a bad light: while some inmates testified that individual correction officers were deliberately blind to inmate-inmate violence, others testified that individual guards assisted them by preventing violence or by rescuing them once an attack had begun. As is discussed in Sections IV and V of this opinion, inmate-inmate violence at CIFM is the result of systematic problems and deficiencies, which tend to negate the efforts of many of the frontline correction officers to curb violence.

The following accounts of violence inflicted upon inmate witnesses by other inmates, drawn as examples from the testimony, show the nature of the problem:[13]

a) James Crosby: Crosby was a twenty-eight year old inmate who had been convicted of shoplifting. He was sent to CIFM in June 1986 after violating the terms of his parole, and remained there until September 1986. He was attacked by other inmates twice. In July 1986, in dormitory 9 Lower, after a quarrel over a radio, another inmate attacked Crosby from behind and then slashed his face with what Crosby claimed was a homemade knife made from a disposable razor blade and a toothbrush. The fight was broken up when a correction officer came into the dormitory. The housing captain who investigated the incident eventually concluded that Crosby's injury was caused by the other inmate's ring; Crosby testified that this captain tried to pressure him into stating that there was no knife. See T. 29-38; PX 45. In August 1986, when working in the clothes storage area, an inmate stabbed Crosby with a sharpened metal rod after Crosby told him not to use his drinking cup. Crosby was treated for a puncture wound. T. 45-61; PX 124 (unusual incident 21MM73); PX 48.

b) James Kenny: Kenny was a twenty-two year old inmate who was sent to CIFM on a forgery conviction early in 1986 and was still incarcerated there when he testified in October 1986. Kenny, who is homosexual, testified to numerous incidents of violence, including the following: First, while in new admissions dormitory 4 Upper, he was threatened with sexual abuse. T. 268-269. Second, while working in the mess hall, several inmates forced him into a utility closet and attempted to rape him; he did not report the incident because he was afraid of reprisal from the other inmates, T. 171-74. Third, while in dormitory 6 Main, an inmate hit Kenny in the face; when Kenny fought back at least three other inmates joined the attack against him. Eventually other inmates broke up the fight, but no correction officer ever intervened. T. 145-150; PX 34. Fourth, in cell area 2 Main, in the course of twenty-four hours, Kenny was sexually harassed, his cell was set on fire, and he was hit on the head and broke a tooth. T. 138-144; PX 16. Fifth, while in the East Module protective custody dormitory, Kenny was involved in four other fights and injured in at least two. T. 154-58, 175-82; PX 15; PX 35.

c) Rene Lisojo: Rene Lisojo, who was eighteen at the time of his testimony in November 1986, was incarcerated at CIFM from March through July 1986 on a grand larceny conviction. He testified to the following incidents: First, in June or July 1986, an inmate chased him with a broomstick *1526 when Lisojo refused to give him his sneakers, tried to extort money and commissary items from him and scraped him with a razor. T. 508-511, 520-21. Second, on the night of June 20, 1986, Lisojo woke up to discover that someone had wrapped burning toilet paper around his hand. He suffered second degree burns. T. 512-13; PX 26. Third, on another occasion, an inmate punched Lisojo in the face in the bathroom, allegedly in the presence of a correction officer. Lisojo required five stitches in the mouth. PX 17. Lisojo testified that the inmate and his friends threatened to slash him if he reported the incident to the authorities, so he told the officer to report that he received his injuries by falling down in the bathroom. T. 514-520. Officer Eder, the officer on duty, denied being in the bathroom and witnessing the punch and testified that he had believed Lisojo's statement that he had slipped and fallen. However, Eder stated that he made no effort to go into the bathroom to determine independently what had occurred. T. 2694-99, 2703-12.

d) Michael Palmer: Palmer, who was twenty-four years old when he testified, was incarcerated at CIFM from April through December 1985 for possession of a controlled substance. He testified to the following incidents, among others: First, on July 1, 1985, another inmate attacked him with a dustpan while he was doing exercises, causing various abrasions and contusions. PX 27, 32 (entry for July 1, 1985); T. 569-70. Second, he was attacked by fifteen inmates who snatched a pendant he was wearing, cut him with a knife, kicked and punched him in the head and face, and stomped and kicked his stomach. T. 571-74; PX 126 (unusual incident 19MM35); PX 30. Third, in an apparently related incident, he was threatened several days later by several inmates with "shanks" (homemade knives), but was helped to safety by a guard. T. 569-70. Fourth, after Palmer was transferred to protective custody in unit 2 Main, a number of inmates attacked him in his cell after they managed to open his locked door. His lock was opened, he believed, by inmates in the officer's section, an occurrence which he had witnessed before. During the ensuing fight, an inmate holding a shank punched Palmer in the mouth. Eventually, Palmer was able to close his door and an officer arrived on the scene. T. 577-585; PX 31.

e) John Rizzi: Rizzi, who was nineteen years old when he testified, was incarcerated at CIFM from May until August 1986, after violating probation on a grand larceny conviction by missing appointments with his probation officer. He testified to the following incidents: On his first day at CIFM, he was attacked in the crowded receiving room bullpen. An inmate stole his jacket and stomped on his hand. Rizzi's hand later became swollen and infected and he had to be hospitalized for a week. T. 665-669; PX 54. Second, a few weeks later, in dormitory 6 Upper, an inmate attacked Rizzi after accusing him of stealing cookies, giving him a black eye, cut lip and bloody nose. T. 669-670; PX 55-56. Third, in dormitory 9 Upper, an inmate punched Rizzi when he refused to agree to give up his sneakers. T. 671-672. Fourth, in work dormitory 12 Upper, inmates tried to extort commissary items from Rizzi and an inmate pulled a homemade knife on him. T. 684-87.

f) Raymond Castro: Castro, who was twenty-one at the time of the trial, testified that he had been convicted of three grand larcenies and a felony charge for sale of a controlled substance. Castro testified about the following incidents that occurred during his second incarceration at CIFM, from the end of 1985 to the beginning of 1986: First, in dormitory 12 Upper and later in 7 Upper, a group of inmates jumped him and kicked him because they said he had pubic lice. T. 857-60; PX 5. Second, in 8 Upper, a group of inmates who accused Castro of being an informant attacked him: they put a blanket over his head, kicked, punched and cut him. T. 861-64; PX 6. Third, in 9 Upper, while watching television in the dayroom, he was jumped by a group of inmates, who kicked him, punched him and swung a broom at him. T. 864-67, PX 9. Fourth, in East Module Protective Custody he was attacked, *1527 punched and kicked but was not injured. T. 870-73; PX 8. Fifth, in another dormitory he was burned on his feet while sleeping and suffered second degree burns. T. 874-75; PX 9.

 
B. Severity of Inmate-Inmate Violence at CIFM

Although no inmate has died at CIFM in the last five years as the result of an inmate attack, plaintiffs have established that much of the inmate-inmate violence at CIFM is extremely serious and some of it is indeed savage. For instance, in the month of February 1987, inmate-inmate violence resulted in sixteen slashings, lacerations or stab wounds requiring sutures or emergency room treatment, and thirteen other serious injuries, including a fractured rib and collapsed lung, a fractured jaw, a neck sprain and a loss of consciousness.[14] Plaintiffs also have tabulated 100 incidents from March 1982 to March 1987 in which inmates were set on fire or otherwise burned, often while asleep in the dormitory at night, or were the victim of an attempted burning, although in seventeen of these incidents, no burn injury was documented.[15] Finally, while none of the inmate witnesses testified that they had actually been raped, there was testimony about sexual harassment and attempted rape and there is documentary evidence of a number of inmate rapes and complaints of rape at CIFM.[16]

Many of these incidents involve knives, both the commercial variety and homemade knives called "shanks" or "shivs", as well as razors, broom sticks, ammonia, bleach and miscellaneous blunt and sharp instruments.[17] In 1986, there were over six hundred violent incidents involving weapons and infractions for possession of weapons.[18] Defendants presented evidence at trial that they had increased the number of cell searches for weapons in recent months, and that, as a result, the frequency of serious attacks has decreased since October 1986. In comparing March 1987 to August 1986, there does appear to be a slight decrease in the number of violent incidents involving cutting or stabbing weapons; however, the March 1987 list still demonstrates the wide use of an impressive array of dangerous objects and weapons.

There is also considerable evidence of record, through documentary evidence and through the testimony of both officers and inmates, concerning the activities of inmate gangs, referred to as "posses," that prey on weaker inmates. For example, there were at least four separate incidents in early 1986 in which inmates asked to be moved to protective custody after being injured or threatened by a posse led by an inmate named James Corey.[19] There was *1528 also testimony about other organized activity such as 1) extortion of money in exchange for sleeping space, called "paying rent," 2) forcing other inmates to wash clothes or do other menial chores for posse members, called "maytagging," and 3) forcing inmates to pay for telephone use, called "running the phone." In addition, many inmates referred to the phrase "snitches get stitches" as a slogan frequently used by assaultive inmates to threaten inmates who complain to the authorities. Defendants argue that they take reasonable steps to break up such activities and that plaintiffs exaggerate the seriousness of organized inmate violence and extortion. Defendants point out, for instance, that three officers testified that they had never heard complaints about maytagging and two had never even heard the term. It is not unexpected, however, that inmates may be more knowledgeable than officers on the types of injuries inflicted upon them by other inmates. I conclude that the evidence establishes that the type of inmate activity described in this paragraph is a significant aspect of inmate-inmate violence at CIFM.

 
C. Defendants' Failure to Control Inmate-Inmate Violence

In addition to the sheer quantity and severity of inmate-inmate violent incidents, the pervasiveness of inmate-inmate violence is demonstrated by evidence establishing defendants' failure to protect inmates from assaults even where a specific need or danger has been identified. This failure is illustrated by the patterns of 1) violence by inmates who have repeatedly attacked other inmates but are permitted to continue their assaultive behavior and 2) violence in protective custody housing units.

1. Repeat Offenders: In PX 376, plaintiffs have compiled documentation about seven inmates who were repeatedly involved in violent incidents, many involving the use or possession of weapons. Their history at CIFM shows that, even when assaultive inmates become known to the authorities for their violent behavior towards other inmates, they are allowed to remain in open general population dormitories. If placed in segregation, which is rare, they are allowed to return to general population after a very short time. Finally, when they are released from CIFM and then return on subsequent convictions, they are placed once more into general population dormitories. For instance, inmate Marco Reyes was placed in administrative segregation from May to November 1986, during which period he was involved in thirteen reported instances of violent or disruptive behavior, including attacks on inmates with a razor, a broomstick, and a pick. Despite this history, in early December 1986 he was moved back to an open general population dormitory, where, within the course of a month, he slashed two inmates with razors. See PX 376.

2. Violence in Protective Custody: Prior to 1985, defendants' protective custody policy was somewhat restrictive, in part because there was often no room to house inmates who required or requested protective custody. In late 1985, however, defendants adopted a policy under which any inmate who requested protective custody status would receive it unless it was known that he intended to do harm to others. To create more room, defendants designated part of the East Module to be an adolescent protective custody area; later, the adolescent protective custody unit was moved to the adolescent new admissions dormitory, without physical separation from the other inmates in that dormitory. T. 2008 (Captain DeCicco). Adult protective custody was also moved back and forth from dormitory to dormitory.

Plaintiffs' expert James Shoultz testified that dormitories are unsuitable for housing protective custody inmates because of their special problems and needs: some need protection and others can be "explosively violent." T. 3810, 3808-10; see also T. 4000-03 (Bair). While Shoultz testified that he had never seen another jail or prison, other than CIFM, where protective custody inmates are kept in dormitories, T. 3809, defendants' expert Chief Painter testified that such inmates may be housed safely in dormitories, and that he used to *1529 do so in his Los Angeles system. T. 4494-95, 4663. (Currently, protective custody inmates in Los Angeles are housed in multiple occupant cells. T. 4495.) Regardless, however, whether protective custody inmates under some circumstances may be housed safely in dormitories, it is apparent from the record of this case that protective custody dormitories at CIFM have not provided adequate protection. PX 374 summarizes thirty-four incidents of fights, assaults, threats, robberies and weapons findings involving East Module protective custody inmates from December 1, 1986 to January 16, 1987, ranging from fist fights to major incidents with substantial injury. The pattern of violence continued after the protective custody unit was moved from the East Module to 7 Upper and 8 Upper in mid-January 1987. In February 1987, twenty-one out of 126 reported incidents of violence at CIFM involved inmates from 7 Upper or 8 Upper, a ratio disproportionate to the ratio of the 7 Upper/8 Upper population compared to the institution's total population. See PX 403 (injury reports for 2/87); PX 263A. These incidents included gang attacks, assaults on sleeping inmates, use of weapons and the infliction of serious injuries.

 
D. The Amount of Inmate-Inmate Violence

Although defendants did not begin to keep records of the total number of violent incidents occurring at CIFM until the trial started, plaintiffs have constructed their own records from injury to inmate reports, clinic injury logs and infraction logs produced to them by defendants.[20] These records document increasing numbers of violent incidents at CIFM, starting with just under 900 in 1982 and rising to over 1300 in 1986, with almost 350 in the first three months of 1987.[21] Plaintiffs also argue that there is a substantial amount of undocumented violence for two reasons. First, plaintiffs point to evidence that defendants regularly lose records of inmate violence, noting that the percentage of missing injury reports, as calculated from the clinic injury report logs, has gone up from eleven percent in 1984 to twenty-six percent in January-August 1986. See PX 418 (explaining calculations of missing injury reports). Second, plaintiffs argue that other incidents of violence go unreported or reported as accidents out of fear of reprisal because of the "snitches get stitches" phenomenon described above.[22]

Defendants criticize plaintiffs' estimates of the amount of inmate-inmate violence at CIFM. Most important, they argue that plaintiffs have incorrectly included in their calculations of reported inmate-inmate violence incidents that defendants term "unsubstantiated," such as claims of assault with a weapon where no weapon was found after a search, incidents of "horseplay," and incidents of inmate violence which defendants argue could not be investigated because they were not reported to the authorities until too long after the fact. Defendants calculate that, of the incidents listed by the plaintiffs from December 1986 to March 1987, twenty-five percent did not result in documented injury. Of the rest, defendants calculate that over half resulted in injuries characterized as "minor," such as bruises, cuts, scratches, swelling or *1530 tenderness, and eighteen percent resulted in "serious" injury.[23]

While these statistics do add perspective to an understanding of the impact of inmate-inmate violence at CIFM, it cannot be concluded that they provide an adequate defense or rebuttal to plaintiffs' claims. First, even eliminating the twenty-five percent of incidents with no reported injury, plaintiffs have still established that there is a large amount of violence at CIFM. Furthermore, although "unsubstantiated" incidents may not merit the same weight as incidents where, for instance, use of a weapon is confirmed, it cannot be expected that in the circumstances at CIFM every weapon will be found, and the heavy volume of such incidents at CIFM is itself a significant indication of the scope of the problem. Finally, incidents such as fights that result in either minor injury or no detectible injury are not factually or legally irrelevant in this case, which, as discussed in Section V of this opinion, concerns the risk of violence faced by CIFM inmates, as well as the actual rate of violence. Even minor fights and horseplay have the potential to explode into violence under the crowded and flammable conditions at CIFM.[24]

Defendants also offered DX VVVV, labelled "Violence Tracking," to show what appears to be a slight reduction in violence from January-February 1986 to January-February 1987, based on the number of inmate infractions during this period. However, plaintiffs point to significant omissions in this chart and, more important, argue convincingly that infractions are not a reliable indicator of violence level because of the large number of cases in which infractions cannot be filed because the assailant is unknown.

In sum, plaintiffs' estimates of the amount of inmate-inmate violence at CIFM represent an impressive effort to reconstruct accurately a complex history of events and are as dependable as can be reasonably expected. It is impossible to make scientifically exact calculations in these situations, and any overinclusiveness on plaintiffs' part is balanced by the phenomena of lost records on the part of CIFM and fear of reprisal on the part of inmates.[25]

 
E. CIFM's Inmate-Inmate Violence Compared with Other Jails and Prisons

Evidence presented at trial, chiefly through the testimony of both plaintiffs' and defendants' experts, established that the rate of violence at CIFM is substantially greater than that in comparable jails and prisons throughout the United States.

Plaintiffs' expert Toni Bair testified that CIFM's rate of violence "far exceeds ... the standards that professional administrators have adhered to," basing his view on "[t]he seriousness of the incidents. The amount of weapons that are used. The serious injuries that the inmates receive from the violence, the frequency of the assaults." T. 3969-70. Bair testified that he had compared data on inmate assaults and fighting in the Virginia prison system, of which he is a regional administrator, for fiscal year 1986 with similar data at CIFM from January-August 1986. Based on his review, he concluded that inmate violence *1531 at CIFM was approximately nine times as high as it was in the Virginia system,[26] T. 3970, even though Virginia's population is all-felon, with more serious criminal histories than CIFM inmates, T. 3976-77, only a quarter of whom have felony convictions, PX 96-98.

Plaintiffs' expert James Shoultz testified that the amount of inmate-inmate violence at CIFM was "very significant," focussing on the number of incidents "that actually involved physical damage to people" as well as the number of incidents involving use of weapons. T. 3776-77. Shoultz estimated that the amount of violence at CIFM was eighty to eighty-five percent higher than would be expected given the nature of the population, which he concluded, after a review of a profile of CIFM inmates, was "very little" different from jail populations in other states, including Florida, where he has been director of corrections in several counties. T. 3777-78. In the Orange County, Florida jail which Shoultz directed from 1972-1981, with an inmate population of around 1000, there were only four or five violent incidents a month involving injury to inmates. T. 3779.

Plaintiffs' expert Verne Cox found a "relatively elevated level of violence" at CIFM, which he had expected based on CIFM's extensive use of large, open dormitories and its high population levels. T. 4726. Cox compared data on violence at CIFM to similar data he was provided in another litigation, concerning Fishkill Correctional Facility, a medium security New York state prison, in which he testified as an expert. Cox, relying on the Fishkill superintendent's testimony about the reporting standard and on Cox' own inspection of the underlying reports, concluded that at Fishkill, any incident of inmate violence which involved "an injury of consequence" was reported as an unusual incident, including "relatively minor abrasions all the way through lethal injuries," T. 4732-33. Hence, Cox compared the number of incidents at CIFM where injury was documented with the total number of reported inmate assaults or altercations at Fishkill. On that basis, he concluded that CIFM's rate of inmate violence was eight to twelve times as high as that at Fishkill. See PX 425 (Cox comparative calculations).

Based on a review of statistics, other documents and testimony, Cox also compared inmate-inmate violence at CIFM to inmate-inmate violence at the Los Angeles County jail system directed by defendants' expert Chief Painter. He concluded that the rate of inmate-inmate violence at CIFM is over three times the rate of inmate violence at the Los Angeles County jail system. However, in at least one barracks-type Los Angeles jail, where inmates are housed in large dormitories similar to those at CIFM, the rate of violence approached that of CIFM.[27] Defendants argue that no comparison is possible between CIFM and the Los Angeles jails, because of differences in physical structure and inmate profiles, and because CIFM reports incidents in which no injury occurs, whereas in Los Angeles, an inmate fight is not reported unless there is serious injury. However, on cross-examination, Painter acknowledged that the definition of inmate-inmate assaults in Los Angeles can extend to include scuffles and pushing matches as well as "a legitimate fight where blows were struck." T. at 4620-21.

Chief Painter himself testified that violence at CIFM is "higher than it needs to be [and] certainly higher than I am used to seeing in other institutions of that size ...," T. 4576, although he added that he thought changes recently instituted at CIFM were starting to reduce the level of violence, T. 4577-78.

Finally, Commissioner Koehler testified that it was his impression that the rate of inmate-inmate violence was higher at CIFM than at the Manhattan House of Detention and other celled borough houses of detention, even though pre-trial detainees are *1532 more difficult to manage and control than sentenced inmates. T. 3560-63. A July 1985 study by the National Institute of Corrections found that, during a one year period, there were no assaults with weapons or sexual assaults at the Manhattan House of Detention and a total of only twenty-one contraband weapons, at maximum, were found. PX 250 at 83.

Defendants argue that none of plaintiffs' experts had sufficient basis to make the comparisons of violence to which they testified, because of their unfamiliarity with CIFM, differences in reporting systems at the different institutions and the fact that none of plaintiffs' correctional experts had directed an institution as large as CIFM. Defendants' specific objections vary in strength. Overall, however, I conclude that all of plaintiffs' experts were amply qualified to make the comparisons to which they testified and gave reasoned bases for their conclusions. Of course it is difficult to compare different institutions with different populations and reporting systems in different parts of the country. But these comparisons, even allowing for variations in reporting systems, still show that the rate of inmate-inmate violence at CIFM is significantly higher than that at other basically similar institutions. The amount of violence at CIFM is also significantly higher than Professor Cox and three seasoned correctional experts Bair, Shoultz and Painter would have expected.

 
III. Staff-Inmate Violence at CIFM: The Problem of Excessive Force

Plaintiffs have established that the use of excessive force by staff against inmates at CIFM is significant and widespread. CIFM is certainly not a jail where sadistic guards regularly torture inmates without cause, and clearly many of the guards who testified at trial appeared to be decent, well-meaning people. However, the evidence does demonstrate that CIFM is an institution where guards, because they are often undertrained, overworked and frustrated, resort too often to excessive and unnecessary force in dealings with inmates, often causing substantial injuries. Even when these incidents are prompted by an inmate's goading behavior, the use of force is often out of proportion to the force required, if any. Even more disturbing, testimony was heard about incidents where there appeared to be absolutely no reason for the use of force on a particular inmate, except that he was in the wrong place at the wrong time.

 
A. Eyewitness Testimony

Fourteen inmate witnesses testified about several incidents of excessive force which the witness either experienced or witnessed. What follows summarizes only a sampling of the credible record in this case on excessive force. As a preliminary issue, defendants make the general point that the credibility of all the inmate witnesses is impaired because each of them had been convicted of crimes, whereas none of the defendants' witnesses had. While inmate testimony about their keepers clearly may be prone to bias and exaggeration, and some of plaintiffs' witnesses were less credible than others, much of the inmate testimony on this topic was credible and sincere, and supported in significant part by documentary evidence. Hence, no blanket finding of incredibility can be made.[28]

The evidence concerning excessive force has been examined with the goal of determining whether plaintiffs have established a pattern and practice of excessive force at CIFM. Each incident described below constitutes a credible instance of misuse of force, and hence a strand of the pattern which I find. However, no attempt has been made to resolve or reach final conclusions as to liability in any particular incident, as would be necessary in the case of an individual claim under 42 U.S.C. § 1983. For instance, a number of the incidents about which testimony was heard followed this pattern: first, an officer was challenged by an inmate, possibly legitimizing *1533 the initial use of a controlled amount of force, but then the incident deteriorated into the unnecessary beating and kicking of an already subdued inmate. In such instances, disputes as to the exact circumstances which gave rise to the beating, or the exact details of the beating, are not as important as the extent of the beating and the extent of the injury.

a) James Kenny: Kenny testified as to numerous incidents of misuse of force against him, the most serious of which occurred in August 1986. At that time, the inmates of dormitory 8 Upper were required to line up after some were accused of misconduct. Kenny asked Officer Tilleli why they had to continue to stand; Officer Tilleli then asked Kenny for his ID card, and Kenny replied that he did not have it with him. Kenny testified that several officers then grabbed him, pushed him against a rail, and told him to go beyond the B gate. When he refused, the officers dragged him out of the dormitory, one by his feet and another by his neck. Then they began punching him; one kept punching him in the side while Tilleli punched him in the face. They handcuffed him, took him into the next dormitory, continued to hit him, threw him on the floor and kicked him. Kenny reported that as a result of this beating, his right thumb and elbow were sprained, his groin was swollen, and his face was lumpy and scratched. T. 247-54; PX 10. He was given an infraction for assaulting an officer, but found guilty only of refusing to produce his ID card; the assault charge was dismissed based on "inconsistent reports on how and where Co. Tilleli was punched." PX 11; PX 152.

Defendants called no witnesses to refute this testimony. However, they point out that on cross-examination Kenny stated that he didn't complain about any groin injury immediately after the incident, T. 301, and point to the investigating officer's statement in the unusual incident that Kenny pushed the officer first and that the inmates disliked Tilleli, so it would be "ridiculous" to credit the inmate allegations against him. PX 152 (unusual incident report 21MM89). Regardless of who pushed whom first, this incident can reasonably be considered as one where an initial use of force spiralled out of control, causing fairly extensive injury, in a situation that could have been handled without any use of force.

b) Rory Hartley: Inmate Rory Hartley was convicted of possession of controlled substances. At the time of his testimony at trial, he had been incarcerated at CIFM for over a year. Hartley testified to the following incident, among others: On February 22, 1986, Hartley was assigned to night sanitation duty. While he was working outside the dormitory the inmates in his dormitory, 5 Lower, were warned that if the inmates continued to make noise, the riot squad would be called. Hartley arrived back at 5 Lower at about 3:10 a.m. and went to bed. Almost immediately, he felt the covers being pulled off him, and saw Captain Chesaniuk, Assistant Deputy Warden DeCanditis and ten correction officers in the dormitory, throwing inmates out of bed and ordering them to the front of the dormitory. As Hartley walked to the front, Officer Henry hit him in the face, Officer Bland grabbed him from the back, and Officer Fisher punched him in the back of the head and kicked him. Fisher also hit two other inmates. When Hartley got to the front of the dormitory, he asked Captain Chesaniuk if he could go to the infirmary. Chesaniuk said that if Hartley would state that he received his injury from a fall, he would not get an infraction. Hartley refused to do this and was given an infraction for disobeying a direct order to stop talking after lights out. The infraction was dismissed at the disciplinary hearing on the basis of the testimony of Hartley and two other inmate witnesses. T. 206-215; PX 20-21. Hartley's medical records show that he sustained a laceration of the right cheek and was given three sutures. He went to Kings County Hospital for x-rays and was diagnosed as having a fracture of the right cheek. He had persistent pain and blurry vision and spent a week in the infirmary. PX 23. Defendants called no witnesses to refute this testimony.

*1534 CIFM personnel failed to report this occurrence as an unusual incident. See PX 158 (unusual incident reports for February 1986); PX 262 (24-hour reports for January-August 1986). Hartley, however, complained to the Warden and the Inspector General. As a result, several of the officers involved were charged with excessive force, making false statements and other violations. As of the completion of the Fisher trial, almost a year and a half after the incident, no action had been taken on these charges. Chesaniuk, who was charged with submitting a false report, failure to investigate an incident and failure to discharge his duties, was promoted a few months after the incident to Captain of Security. T. 2939-40 (Garvey). On July 22, 1987, Chesaniuk pled guilty to failure to investigate and nolo contendere to the other charges: he was given probationary status for one year and forfeited ten days vacation. On July 29, 1987, DeCanditis pled guilty, was placed on probationary status for six months and forfeited thirty vacation days. Bland resigned on July 15, 1987, and the charges were administratively filed. Koehler Supp. Aff. at ¶ 8(d). Officer Fisher was never charged in this incident, although defendants argue that this was appropriate because his participation was never sufficiently established.

This incident is a shocking example of excessive force which seems to have been inflicted on Hartley with no justification, as if at random.

c) Keith Beattie: Keith Beattie has been incarcerated at CIFM twice; at the time of his testimony he had served about five months of a sentence for grand larceny. Beattie testified to the following incident that occurred on August 15, 1986. According to Beattie, he was working at his job collecting garbage cans, when Officer Young of dormitory 12 Main became angry at him for failing to take out his dormitory's garbage. Young and three other officers gathered around Beattie, backing him against the wall. Beattie sat down on the floor to avoid conflict. Young then threw the trash can on him and ordered Beattie to pick up the garbage. Beattie refused and stood up. Young and Officer Gaylord then hit Beattie and knocked him down. Young punched Beattie again in the eye and stomped on Beattie's stomach. Beattie was then ordered to pick up the garbage. After doing so, Beattie bolted down the corridor. Two officers chased him to the next gate and forced him back, punching him. Young then said, "I'm not finished with him," and Gaylord kicked him in the stomach. Beattie fell to the ground, having trouble breathing. Twenty minutes later, Officer Cubano arrived and sent Beattie to the clinic. T. 330-342. The clinic reports document that Beattie sustained swelling to his left forehead, an irritated left eye and abdominal sounds. PX 51.

Officer Young testified at trial. He stated that he confronted Beattie because his dormitory's garbage can had not been emptied. He stated that Beattie swung at him, he tackled Beattie, and they both fell to the ground. T. 1286-90. He testified that he punched Beattie while they were both on the ground, and that "of course" Beattie physically resisted before he punched him. In his deposition, however, Young testified that neither he nor Beattie did "anything physical" after being brought to the floor. T. 1370-71. Young denied that any other officer hit Beattie.

Officer Arenella and Captain Cubano also testified about this incident. Arenella testified that he did not witness or hear the argument leading to the fight, but that he heard "scuffling of feet" from his post, which was about thirty feet away, T. 1292, saw Young on the floor with Beattie, separated them, and then returned to his post. T. 1472, 1479. Captain Cubano testified about the investigation that he conducted, in which he concluded that the force used by Beattie was "necessary and minimal," PX 51. Cubano stated that he stayed at least five hours after his second tour of duty to investigate the incident. However, in some respects the investigation was not thorough. Cubano made no inquiry as to the participation of other officers in the incident alleged by Beattie, because, he explained, Beattie did not raise this allegation until about an hour after his trip to the clinic, when he submitted his written statement. *1535 Cubano did question some inmates in the dormitory, and received a written statement, allegedly from inmate Godfrey Frith, which states that an unidentified inmate swung at "my CO." The handwriting on the report shifts midway from an illiterate style to one very similar to Officer Young's. PX 51; T. 1384-87 (Young); T. 2447-48 (Cubano).

Beattie struck the court as a hot-tempered character but his testimony seemed sincere. Young's testimony also appeared to be credible in large part, but aspects of his story were confused, and the change in handwriting on the inmate statement was striking. In sum, while this incident was difficult to assess, it appears to have been one in which Beattie was subjected to excessive force by one or more officers. While Beattie's provocative behavior may have goaded Officer Young, that need not have led to an incident of this nature.

d) Karen DeCarlos: Inmate Karen DeCarlos, who is male, has been incarcerated numerous times at CIFM. At the time of his testimony, he was being held as a parole violator on prostitution charges. DeCarlos testified that he had been assaulted by CIFM officers five to seven times. In one of these, Captain Warshawsky came to the North Module to give DeCarlos an infraction and to take him to the clinic. DeCarlos said that nothing was wrong with him and that he would not go. Warshawsky grabbed him by his left armpit and squeezed it hard enough to leave fingernail marks. DeCarlos cursed at him and tried to pull away. Warshawsky then pushed him down, kneed him in the face, blacked his eye and busted his lip. T. 390-92; PX 24. Defendants, however, point to evidence that DeCarlos refused medical treatment after the incident and didn't report the black eye until 24 hours after the incident. DX P, Q; PX 24. Warshawsky testified that, as he was taking DeCarlos to the clinic, DeCarlos pushed him in the chest and "took a fighting stance." T. 2068. Warshawsky then punched him twice in the head. T. 2093-94. Although he is 6'3" and DeCarlos is 5'2", Warshawsky indicated that there were no alternative methods to restrain DeCarlos. T. 2071-72. Warshawsky was given a corrective interview for failing to exercise options when dealing with a recalcitrant inmate. He testified that in retrospect he would not have handled this situation differently. T. 2073.

e) Michael Palmer: Palmer testified that on October 29, 1985, an officer attacked him with a riot stick and broke his arm. Palmer had been sent to Kings County Hospital for tests for a persistent wrist complaint. On returning to CIFM at about 11:00 p.m., he was placed with new admissions in the receiving room and complained about this. Officer Tilleli called him a "wise guy" and told him to "come out of the bullpen, he got something for [him]." T. 547. Palmer refused, and Tilleli came back with a riot stick and hit Palmer on his left knee. Officer Fisher then ran into the pen and hit Palmer in the chest several times. Palmer fell and Tilleli hit him several more times with the stick. He tried to hit Palmer in the head but Palmer blocked the blow with his arm. Finally, Officer Long told the officers to stop. Then Officers Tilleli and Fisher dragged Palmer to an empty bullpen, where Officer Tilleli kicked him in the side several times. Eventually, Palmer was taken to the clinic and found to have a broken arm. T. 544-57. An unusual incident report was filed concluding that Officer Tilleli used unnecessary force and poor judgment and recommended disciplinary charges against him for unnecessary force. PX 162 (report 18MM148). No action was taken on these charges, however, and Tilleli resigned in fall of 1986. DX XXX at ¶ 11. Palmer was given an infraction for assaulting an officer, but the hearing officer found no evidence to support the charge and dismissed it. PX 29.

Defendants called no witnesses to refute Palmer's testimony. They point out, however, that Palmer testified on direct examination that he had only three prior convictions, while on cross-examination he admitted to eight prior convictions, arguing that this casts doubt on the credibility of Palmer's account. In addition, defendants argue that the broken arm could have been the result of a prior injury, that Palmer lied *1536 about the repetitious blows and kicks, and that there is no evidence to confirm that Fisher was involved or that Long witnessed the incident. However, Palmer's account was fundamentally credible overall and was even supported in significant part by defendants' own investigation and conclusion that Officer Tilleli had used unnecessary force.

f) Roger Ramsey: Ramsey was convicted of possession of a weapon and possession of fraudulent instruments. He was incarcerated at CIFM from September 1985 to June 1986. He testified about witnessing several incidents of excessive force that occurred to other inmates and recounted the following incident that happened to him: On November 12, 1985, at 4:30 p.m., Ramsey was in his dormitory, 6 Lower, waiting for permission to go to school. Officer Yvette Johnson was on duty. Ramsey heard the phone in the officer's station ring and heard Officer Johnson tell the person that there was no problem in the dormitory and that the alarm must have gone off by mistake. Then Ramsey saw ten to fifteen officers outside the gate responding to what they thought was an alarm. Officer Johnson stepped out of her station, saying "no, its a false alarm," but the officers pushed past her and rushed into the dormitory. Ramsey panicked, ran into the dormitory and backed himself onto the lower bunk of a bunk bed. The officers jumped on Ramsey and began hitting and kicking him. Six officers hit him, but he only recognized one, Officer Vanderpool. Vanderpool grabbed Ramsey by the collar, snatched him off the bunk, swung him against the iron rail on the bed and punched him in the face. Finally, Captain DeCicco ordered the officers to stop. T. 935-43. Ramsey's medical report confirms his injuries, including swelling and tenderness of the left face, laceration of the left upper lip, abrasions on the lower back, pain and tenderness of the thigh. He was given four sutures for the laceration, an ice pack and motrin. PX 14.

The facility investigation led to an unusual incident report that failed to resolve how Ramsey received his injuries or the extent of the force used against him. PX 13. Defendants argue that the evidence indicates that Ramsey attempted to hide under the bunk bed, not on the lower bunk, and that his injuries occurred when he was pulled out from under the bunk, rather than from a beating. Ramsey's account was credible, however, and his injuries consistent with punches to the face. Defendants also point out that Vanderpool later resigned, DX XXX at ¶ 11, and that, after this incident, mock alarm drills were held to train the officers on proper procedures.

g) The October 1986 Disturbances: Jose Vasquez and Ross Bennett: In October 1986 Vasquez and Bennett were both "state inmates" who were held at CIFM on parole violations, as well as on pending charges. They testified about the infamous incidents that occurred at CIFM and elsewhere on Rikers Island from October 13 through October 17, 1986. The April 1987 report of the State Commission of Correction, Inquiry into Disturbances on Rikers Island, October 1986, is also in evidence as PX 95.

Testimony from Vasquez and documentary evidence described the following events: On Monday, October 13th, the inmates in 5 Main began to act up, breaking up beds. T. 3631-41. The inmate dispute was resolved peacefully later in the day with the signature of a memorandum of understanding concerning several aspects of treatment of state inmates at CIFM. T. 3645-50, PX 19, PX 89. On October 14th, the officers on the 4:00 p.m. tour refused to report to their posts "as a result of existing poor working conditions and prior unattended grievances," PX 94; PX 95 at 71-72. The matter was not resolved until midnight and the 8:00 a.m. to 4:00 p.m. shift had to be held overtime. Id. On October 16th, parole violators in the East Module barricaded their housing unit because of grievances and negotiations were held. PX 93; PX 95 at 73-74.

On Friday, October 17th, there was a disturbance in the CIFM mess hall, which began when an inmate and an officer got into a verbal dispute. Although Vasquez testified that the inmate offered no physical resistance, the situation deteriorated *1537 into violence, and the inmate was beaten up by numerous officers. Other inmates were beaten too; Vasquez was beaten as he tried to leave the mess hall, and was further beaten as he waited outside the clinic for treatment. He sustained two black eyes, a swollen face, abrasions to his back and shoulders and lacerations of the head and wrist requiring stitches. T. 3651-3666; PX 91. He was never given any infraction or criminal charge in connection with the mess hall incident.[29] Inmate Ross Bennett also witnessed part of this sequence of events: he was outside the mess hall and witnessed another inmate being stomped, kicked and beaten. T. 3727-29. A number of officers were also injured by inmates in this incident.

Bennett testified that, after the mess hall incident, there were continuing threats from officers, along the lines of "it ain't over yet." T. 3729. At 9:30 p.m., a captain and a deputy came to Bennett's dormitory and told the inmates to pack their property. They were then taken to the House of Detention for Men ("HDM") by bus. Outside the bus were two rows of officers, with fourteen officers in each row, dressed in riot gear. The inmates, who were handcuffed to each other, were led into HDM through the two rows of officers, and the officers punched them and struck them with sticks. Bennett was hit in his head, back and thighs. T. 3730-36. The State Commission of Correction concluded that the inmates were forced to run a gauntlet of baton-wielding officers from the CERT team, and that "the use of excessive and unnecessary force during the move from CIFM to HDM" resulted in thirty-four injury reports, twenty-four actual injuries, and eight "serious" injuries, including head and scalp lacerations, one loss of consciousness, contusions, abrasions, bruises, swellings and a possible rib fracture. PX 95 at 83-86.

Defendants argue that this testimony is irrelevant because 1) even plaintiffs do not allege that these events were typical; 2) the running of the gauntlet did not itself occur at CIFM, and plaintiffs did not establish that CIFM officers were involved; 3) to the extent CIFM officers might have been involved, it would have been in their capacity as members of the now-substantially revised Correction Emergency Response Team ("CERT"), and not in their capacity as CIFM staff and 4) at the time the inmates went through the gauntlet they were no longer Fisher class members because they had physically left CIFM. Defendants also argue that they were unable to put on a defense to the testimony of Bennett and Vasquez because, at the time of their testimony, there were ongoing departmental, state and federal investigations, and counsel could not interview or call individual officers as witnesses because of uncertainty as to what charges would be filed against whom.

I conclude, as I stated at trial, that it was appropriate to hear testimony and receive evidence on the October 1986 disturbances, because "the subject matter seems to [be] at the heart of what the whole purpose of this trial is, and to ignore it is unrealistic," T. 3621, and would be irresponsible. Although defendants may have been circumscribed in their response to the inmate testimony by ongoing investigations, it cannot be concluded that they were unfairly prejudiced. Testimony about the gauntlet itself, which took place outside CIFM, should be set aside in deciding the issue of staff-inmate violence. However, it is appropriate to consider the evidence about the other events that took place at CIFM during this time period, and this evidence is consistent with and supports the other evidence about excessive force in this case. Finally, the October 1986 disturbances are an indication of the tensions brewing below the surface at CIFM that could erupt again, if steps are not taken to prevent use of excessive force in the future.

h) Defendants' Witnesses: Plaintiffs argue that some of defendants' own witnesses *1538 supported plaintiffs' claims of misuse of force. Two incidents were brought out that seem especially telling. First, Officer Robert Murray testified that he was involved in a violent incident with inmate John Beatty in June 1986. Murray testified that he and Officer Saglimbene grabbed Beatty and wrestled him to the floor after the inmate "swung his arms at me in a threatening manner." T. 1160. The captain who investigated the incident wrote that, "[i]n conclusion I find that this is another case of lack of common sense. I feel that according to the reports submitted by Officer Murray and Saglimbene the entire situation was avoidable and the assault on staff was actually provoked by staff." He also noted that Officer Murray has "been involved with at least five compensation cases involving the use of force" in the last year. PX 154 (unusual incident report 20MM136) (emphasis in original). The captain recommended that Saglimbene be given a corrective interview on the proper use of force and that Murray be given a corrective interview and placed on probation. The tour commander concluded that "the force used to control this incident could have been avoided, making the force used unnecessary." Id., part B. Murray testified that he was given a corrective interview for failing to report the incident correctly, but the unusual incident report shows that the interview was "for failing to properly supervise inmates while performing their [sic] duties." Id., part C. Murray was not placed on probation.

Second, Officer Jose Jordan testified on direct examination that he was involved in only two incidents involving use of force. On cross-examination, it was brought out that Jordan has been involved in additional uses of force. T. 1444. In one incident, Jordan testified that he was working in the main yard when an inmate swung a base-ball bat at him. He stated that he punched the inmate to avoid being hit, and then the response team took the inmate back to the institution. T. 1417-18. On cross-examination, however, Jordan acknowledged that the inmate was not taken back to the institution directly. Rather, the response team took the inmate to a bus where the inmate was beaten by several officers. T. 1439-42. While Jordan said he could not remember the extent of the inmate's injuries, the records indicate that the inmate had a hematoma over the nasal bridge, contusions and abrasions of the face and hematomas of lips and gums, and a broken nose. PX 167 (unusual incident report 17MM50).

Finally, in fairness to defendants, it must be pointed out that much of the testimony of the correction officers was favorable to them. Many of the officers impressed the court as bright, sincere and competent professionals who handle their difficult charges and their stressful work environment well. In fact, several officers testified that they had never slapped, kicked or punched an inmate, T. 2688 (Eder); T. 1732 (Bocina); T. 1483-84 (Arenella); and others testified to uses of force that were completely appropriate.

In sum, however, the eyewitness testimony heard, bolstered by the voluminous documentary evidence of record, supports a finding that there is a pattern of excessive force at CIFM, manifesting itself, inter alia, in the recurrence of 1) use of force out of frustration in response to offensive but non-dangerous inmate goading; 2) officers' use of excessive force as a means of obtaining obedience and keeping order; 3) force used as a first resort in reaction to any inmate behavior that might possibly be interpreted as aggressive; and 4) serious examples of excessive force by emergency response teams.[30]

 
*1539 B. The Amount of Staff-Inmate Violence at CIFM

It proves difficult to arrive at an accurate assessment of the quantitative amount of staff-inmate violence at CIFM. Defendants have not kept specific tabulations of inmate-officer violence, but plaintiffs have reconstructed estimates of staff-inmate violence from three sources of information: 1) unusual incident reports, which are filed whenever a use of force results in a "medically detectable injury or any degree of visible physical injury, but ... not necessarily a serious injury," PX 236 at 21; 2) use of force reports, which are filed when force is used without regard to degree of injury or other circumstances, PX 221 at 3; and 3) injury to inmate reports, which are customarily filed whenever an inmate is sent to the clinic to be examined for an injury or possible injury, T. 1141 (Officer Murray). Using the unusual incident reports and use of force reports, plaintiffs chart an uneven rise in the number of use of force incidents, from 159 in 1982 to 201 in 1986.[31] However, when plaintiffs examined the injury to inmate reports, they found many more injury to inmate reports reflecting staff-inmate violence than were accounted for in the unusual incident and use of force reports, indicating that the actual amount of use of force is significantly higher.[32] They also cite other deficiencies in record-keeping that tend to indicate underreporting of staff-inmate violence.

Defendants' exhibit VVVV, based on defendants' new system of reporting violence biweekly, indicates that use of force incidents declined fifty-six percent from January-February 1986 to January-February 1987. However, the figures in this exhibit do not correlate with the combination of numbers that can be gathered from the use of force, unusual incident and injury reports for these periods. A further problem with both plaintiffs' and defendants' charts is that they summarize the total of all uses of force and do not separate legitimate from illegitimate uses of force. However, defendants themselves have concluded that force was misused in twenty-five instances in the two years from February 1985 to January 1987.[33] In comparison, in the entire Los Angeles County jail system, with roughly 22,000 inmates, Chief Painter testified that there were only ten or fifteen findings of misuse of force a year. T. 4640. In addition, of course, a core part of plaintiffs' case is evidence that CIFM investigations of misuse of force and discipline of offending officers are so poor that many misuse of force incidents are not reported as such.

Plaintiffs' experts were struck by the prevalence of staff-inmate violence at CIFM. Bair termed the level of officer-inmate violence "high," and stated that violence is "the way of life at CIFM. That's the way staff and inmates resolve issues[:] ... by putting their hands on one another in violence." T. 4026. Shoultz called the amount of violence "rather significant" and "unusual." He counted over sixty incidents of staff-inmate force during the period June-August 1986, T. 3834, and said that he would normally expect two or three such incidents a month per a thousand inmates at a comparable institution. T. 3835. Nathan commented on the "[s]pontaneous, unfettered, uncontrollable venting of anger in some cases" at CIFM. T. 4413. Nathan also described the large amount of "severe contact" between staff and inmates at CIFM, such as punches to face and kicks to the groin: "It is a most unusual phenomenon to me that inmates are continually being punched in the face. It is not simply a control mechanism I have seen in any prison in which I have been involved where force is controlled." T. 4293. He also *1540 noted that "[w]hen inmates live in a lawless environment, they tend to be more lawless, more violent than they otherwise would be," T. 4322.

Defendants' expert Painter called the CIFM officer-inmate violence "higher than it might be" and acknowledged that it was "higher than it needs to be," T. 4585, 4631, although Painter denied that officer-inmate violence was "systematic" in the sense that "all of the officers are using force all of the time." T. 4585. He, like plaintiffs' experts, suggested that force was used as a disciplinary measure: force "is higher than it might be because ... I don't believe they have other tools to deal with ..., for example, a consistent inmate disciplinary process." T. 4585. The officers "want to feel that they have some control, and it is not an issue of force strictly because they want to punch people but because they feel they need to be in control of the institution." T. 4642. Painter was also struck by the prevalence of street fighting techniques used by the officers, stating that he "believe[s] that they're poorly trained in the use of force, ... and they are more likely transferring their street fighting skills into their correction officer job." T. 4584.

In addition to this testimony by experts that use of force at CIFM is unexpectedly and unnecessarily high, there is also some evidence of record comparing the level of use of force at CIFM to that at other prisons and jails, although there is less comprehensive comparative evidence than there is on the issue of inmate-inmate violence. As discussed above, there are roughly as many officially documented instances of misuse of force at CIFM, an institution with roughly 2500-2600 inmates on a daily basis, as in the entire Los Angeles County jail system, with roughly 22,000 inmates. Nathan testified that the amount of staff-inmate violence at CIFM is in the mid-range compared to the staff-inmate violence in the Texas correctional institutions he has studied. T. 4412.[34] In addition, Koehler testified that it is his impression that officers at the Manhattan House of Detention and the other single-celled borough houses use less force than at CIFM despite their more difficult pre-trial detainee population. T. 3560-63.

 
IV. The Roots of Violence at CIFM

The evidence at trial established that five major problem areas at CIFM are significant causes of violence, both inmate-inmate and staff-inmate: 1) overcrowding; 2) excessive reliance on dormitory housing; 3) lack of adequate classification; 4) inadequate staffing and supervision and 5) inadequate systems for controlling, investigating and disciplining staff misuse of force.[35]*1541 While defendants do not concede the connection, their own actions during and after the trial to attempt to reform their practices in many of these areas suggest an implicit recognition of the link between these subjects and violence at CIFM. The changes in defendants' policies and practices that occurred during the course of and after the trial are discussed below where relevant. The legal significance of the changes are treated in connection with the necessity for an injunction and the scope of that injunction in Section V of this opinion.

The five significant areas in which violence at CIFM has its roots will be discussed in some detail, because an understanding of the causes of violence at CIFM is important both in determining liability and in fashioning a remedy. However, it is important to remember that the sole issue tried in this case thus far is violence, and the evidence on overcrowding and the other four areas is relevant only to the extent they are found to contribute to violence at CIFM. As an example, the question is not presented here whether CIFM is unconstitutionally overcrowded: the question rather is whether plaintiffs have sufficiently demonstrated that crowding at CIFM is a cause of violence.

 
A. Overcrowding

The parties spent much time and effort at trial and in their post-trial briefs on the issue whether overcrowding at CIFM is a cause of violence, or, more accurately, the extent to which overcrowding contributes to violence at CIFM. It is important to note, however, that the link between violence and overcrowding in the prison setting is hardly an issue of first impression. Over a decade ago, this court, largely on the basis of testimony from corrections experts and psychologists, determined that overcrowding, among other adverse consequences, intensified inmate hostility and aggression at the Bronx House of Detention. See Ambrose v. Malcolm, 414 F. Supp. 485, 487-93 (S.D.N.Y.1976). Since then, numerous courts have found overcrowding to be a significant cause of violence in various jails and prisons. See, e.g., Mitchell v. Cuomo, 748 F.2d 804, 806-07 (2d Cir. 1984); Toussaint v. Yockey, 722 F.2d 1490, 1492 (9th Cir.1984); Albro v. County of Onondaga, 627 F. Supp. 1280, 1286 (N.D.N.Y.1986); Vazquez v. Gray, 523 F. Supp. 1359, 1361 (S.D.N.Y.1981). The body of knowledge contained in these opinions does not make a finding of a connection between crowding and violence at CIFM a foregone conclusion: clearly, the individual situation and circumstances at CIFM must be taken into account. However, this court does not write on a clean slate.

 
1. Overcrowding at CIFM:

CIFM's official capacity has been listed by defendants as 2083 since early 1985, based on an allocation of sixty square feet per male inmate in dormitory housing set by the City Board of Correction and the State Commission of Correction. PX 263A, 360, 362A. CIFM has reached population peaks as high as 2800. PX 263A. Defendants concede that CIFM regularly operates over its capacity: on April 29, 1987, a peak day in 1987, defendants calculated that the institution was at 135% of its capacity. Id. Plaintiffs argue that in fact CIFM is even more overcrowded than its official capacity suggests, because of restrictions on population in some special purpose dormitories. The average daily population at CIFM during the course of the trial was in the range of 2500-2600. T. 4691 (Cox).

Most CIFM dormitories contain 3456 square feet of sleeping space with 651 square feet of dayroom space. PX 247 (admissions). Thus, when dormitory populations reach 100, each inmate is allotted 34.6 square feet of sleeping space and 6.5 square feet of dayroom space. The dormitories are frequently packed with double bunks. The American Correctional Association ("ACA") Standards for Adult Correctional Institutions and those for Adult Local *1542 Detention Facilities provide for a minimum of fifty square feet of sleeping space in multiple occupancy housing; no more than a total of fifty inmates are to be held in a single multiple occupancy room, and inmates are to be allotted thirty-five square feet of dayroom space. PX 372-73.

Defendants argue that crowding at CIFM is ameliorated by inmate work and recreation activities. In addition, Chief Painter testified that CIFM dormitories have more "undesignated" space than most crowded jails: the "aisles were wider, the spaces between the bed[s] were wider, and the area not occupied by bunks was greater than it is certainly in [his] facilities." T. 4559. However, these features could only partially alleviate the burdens caused by the concededly high level of overcrowding.

 
2. Connection between overcrowding and violence:

The connection between overcrowding and violence was recognized at least to some degree by the corrections experts and research experts on both sides, as well as by many of the Department of Correction personnel who testified.

Speaking from practical experience, several correction officials and CIFM correction officers testified that it was their impression that there is a real connection between prison crowding and prison violence. For instance, Officer Jeffrey Young testified that inmates fight over "[d]isrespect, space and I guess just getting on each other's nerves," T. 1322, explaining:

 
You have double bunks. When you have a hundred inmates in a dorm, especially when things is hot, the weather, space. I mean you feel cluttered up. You have guys, I mean it's just frustrating at times. At nights you have a guy that's sleeping on top of you who doesn't wash as much as you and he might be stinking and you might have an attitude and you feel like he's just on top of you. And one thing might lead to another. Why don't you go take a shower, this and that. And there you go.... A fight might break out.

T. 1340-41. Officer Robert Murray, when asked by the court if he had any suggestions on how to reduce violence at CIFM, answered, "Cut down the population. There's too many inmates in one dormitory." T. 1254. Warden Garvey acknowledged that crowding contributes to jail violence, agreeing that too much unwanted contact at close quarters creates tension which can then turn into violence. T. 2947-48. Commissioner Koehler testified that, while he did not think that crowding was a direct cause of the October 1986 Rikers Island disturbances, "one of the variables ... that contributes to unrest in jail systems is the density of the population...." T. 3455.[36]

Plaintiffs' correctional experts Bair and Shoultz, based on their experience in the field of corrections and on their personal observations of conditions at CIFM during their tours, as well as their familiarity with the record of overcrowding and violence at CIFM in this litigation, both testified that in their judgment overcrowding is a major cause of violence at CIFM. They concluded that even if other problems such as staffing and classification were addressed, violence cannot be controlled unless the issue of overcrowding is addressed. See T. *1543 3978, 4025-27 (Bair); T. 3780, 3835, 3943-44 (Shoultz). Although defendants' expert Chief Painter took the position that overcrowding is not a direct cause of inmate violence, he acknowledged on cross-examination that the Los Angeles County Sheriff's Department Defensive Tactics Manual, which he personally approved, states that an officer's responsibilities to guard the safety and welfare of staff and inmates "becomes more difficult due primarily to the overcrowding of all custodial facilities which has created more dangerous and violent physical hostilities towards deputy personnel, civilians and other inmates." T. 4629.

Although much was made of the differences in views between plaintiffs' research expert Verne Cox[37] and defendants' research expert Gerald Gaes,[38] there was actually not such a substantial disagreement between them on the relationship between crowding and violence. Professor Cox, who has had a long and distinguished career, was one of the original pioneers in the field of psychological research on the effects of crowding in prisons. Dr. Gaes, a younger man who has begun to make a name for himself in this field, is trying to refine and test the limits of Cox's work. This dynamic between the two research experts helps to establish a framework in which to compare their opinions.

Professor Cox testified that the high population density at CIFM causes inmate stress and arousal, manifesting itself in violence as well as other adverse consequences. Cox' theories and research point to a substantial connection between overcrowding and inmate-inmate violence: a high population combined with predominance of open dormitories causes stress and arousal, one of the manifestations of which is violence. T. 4692-94. According to his research, "social density," the number of persons in a particular space, is a more potent factor in creating negative consequences than "spatial density," the actual amount of space per person. T. 4696-97. Cox' research shows that the negative effects of social density include a variety of measures of stress, such as illness complaints, mood states, natural deaths, psychiatric commitments and violence. PX 411. Cox explains these findings in terms of "social interaction demand": crowding produces uncertainty arising from contact with unfamiliar individuals, "goal interference" and resulting frustration largely related to competition for resources, and "cognitive load," which reflects the difficulty of making decisions in a complex situation. The negative effects of social density are magnified in prison because of the relative dangerousness of the environment, the high turnover and constant influx of unfamiliar persons, and the limited resources available to each prisoner. See generally PX 411 (Cox, Paulus, McCain, Prison Crowding Research, reprinted from American Psychologist, October 1984); PX 412 (McCain, Cox, Paulus, The Effect of Prison Crowding on Inmate Behavior).

Cox endorsed the views expressed by Susan Saegert, an expert on the psychological effects of crowding, concerning "social overload" and its negative consequences in prisons, which were largely adopted by this court in 1976 in Ambrose v. Malcolm, 414 F. Supp. at 492. When asked how the state of knowledge on the subject had changed since 1976, Cox stated that "a much more *1544 substantial body of data collected within prisons has emerged that basically buttresses many of the points that Susan Saegert stressed in her article...." T. 4700.

Cox also testified that the short length of stay of inmates at CIFM did not change his view that crowding was a major cause of the high rate of violence at CIFM, because "the other side of that coin is that shorter lengths of stay will result in higher population turnover.... The high turnover places a greater demand on the individual [] having to face unknown people and situations that can pose greater threat and harm." T. 4745. In addition, violence, unlike other manifestations of stress, such as increased natural death rates, does not take a long period of time to develop and, once inflicted, has an immediate effect.

Defendants criticize Cox' qualifications to testify concerning the connection between crowding and violence at CIFM because he has no experience working with inmates or in prison administration, T. 4690, and because he has done little additional new research on crowding since 1982, T. 4776. In addition, Gaes criticized Cox' model as overly theoretical and a poor predictor of what occurs in reality, stating that he didn't "see a whole lot of data consistent with it," T. 4864, and that it failed to take into account issues of resource allocation. T. 4864-65. Gaes also argued that, at any rate, there are significant differences between CIFM, a short-term jail, and the long-term prisons at which Cox studied inmate violence. In an article by Cox' long-time collaborators Paulus and McCain, entitled Crowding in Jails, the authors concluded that the short length of stay at jails could ameliorate the negative effects of crowding. Cox testified that he disagreed with this conclusion. T. 4752.

However, much of Gaes' own research and writing supports Cox' core conclusion that there is a substantial connection between overcrowding and violence in correctional facilities. In his 1985 paper, Prison Violence: The Contribution of Crowding versus Other Determinants of Prison Assault Rates, DX DDDD, Gaes examined data on assaults of all types from nineteen federal prisons for a period of thirty-three months; in addition to crowding, he examined the effects of many other control variables such as prison size and characteristics of prison population. He analyzed this data via a sophisticated multivariate technique that allowed him to assess the effects of all the variables simultaneously. Gaes concluded that "[c]rowding operationalized as the number of inmates to an institution's rated capacity proved to be an important determinant in the level of assault rates for the federal prison system.... The data indicate that in this period assault rates generally increased with crowding increases." DX DDDD at 62. He also found that "a system configured with larger percentages of their population housed in dormitories is especially susceptible to higher assault rates." Id. Gaes reconfirmed these findings in a later review of crowding literature, see PX 414 (Gaes, The Effects of Overcrowding in Prison), in which he concluded that one of the "basic conclusions warranted by the prison crowding research" is that "prisons that have higher density ratios are also more likely to have higher assault or misconduct rates," id. at 136. He noted elsewhere that "a prison with an excessive number of inmates housed mostly in dormitories is particularly likely to have higher assault rates," and that "[i]n summary, crowding, not age or transiency, is the best predictor of assault rates." Id. at 134.

Gaes' central thesis and the core of what he views as his disagreement with Cox is that the relationship between overcrowding and violence is relatively inelastic, and that it is important to control for the effect of other variables carefully. Hence, Gaes testified that the relationship between crowding and violence "is rather weak. So that it takes a rather large change in social density to produce a rather small change in the assault rate." T. 4852. He testified that his research has led him to conclude that a 100% increase in density at a correctional facility would yield a change of only one-half an assault per month per thousand inmates. T. 4854-55. However, such a rate of change is not *1545 insignificant. In fact, Dr. Gaes' results also show that a 100% increase in density would yield an increase of 159% in the number of assaults. Gaes assumed a basic assault rate of 8.5 assaults a month per 10,000 inmates. Taking a correctional facility with a population of 10,000 and increasing the population by 100% to 20,000 would lead to a total of seventeen assaults per month, even before taking into account any increase in the assault rate. The increased assault rate of half an assault per one thousand inmates would lead to an additional five assaults a month, bringing the institution to a total of twenty-two assaults per month. This represents an absolute increase of 13.5 assaults per month, an 159% increase from the base rate of 8.5. T. 4938-41 (Gaes cross-examination).[39]

Finally, Gaes concluded his testimony by warning that models and predictors of the relationship between crowding and violence should not be swallowed whole: "I am showing you that there is much inconsistency and that you have to very carefully weigh what has been found in the literature against what your intuition is and what your findings are at C76 itself, because there are so many differences between different systems." T. 4974. Gaes has also expressed in his work a philosophical or political warning that:

 
There are issues that prison crowding research can and cannot address. Optimistically, prison crowding research can define the parameters of prison crowding (when and if crowding effects occur), the extent of debilitation, and the role of intervening variables that may allow intervention in the crowding, debilitation process. These data could be used to examine individual crowding suits and help administrators set standards. What prison crowding research cannot do is distinguish levels of risk associated with imprisonment from levels of risk that are considered cruel and unusual punishment. This latter determination falls under the auspices of the courts, the influence of legislators, and the conventional wisdom and morality of the community.

PX 414 at 141. While both warnings are well-taken, neither casts significant doubt on Cox' central thesis or the practical reality attested to by CIFM personnel that over-crowding is a major cause of violence at CIFM.

One more significant aspect of Gaes' testimony on overcrowding must be addressed. Gaes attempted to graph the relationship between changes in CIFM's population and the rate of violent incidents from January 1985 to January 1987. DX PPPP. He testified on direct examination that he could find no statistical relationship between population and infraction rate, noting that the infraction rate sometimes followed population fluctuations and sometimes diverged from them. On cross-examination, Gaes stated, however, that much of the apparent divergence occurred at times when the figures for population, infractions or both were unreliable or were computed in ways that were not comparable. In fact, Gaes acknowledged that the infraction rate did follow the population level roughly for about thirty-five of the forty-four months, *1546 with two interruptions, one of six months and one of three months. T. 5011. During the first, from July to December 1985, the major divergence took place after October 1985, when CIFM's warden died and a new warden took his place. The three month interruption occurred during the period from August through October 1986: in August 1986 there was an inmate insurrection in the East Module and in October 1986 the disturbances at CIFM described earlier took place. Another apparent divergence is caused by an error: the graph indicates that the population in January 1985 reached 2700, a sharp rise in population not accompanied by a rise in infractions, but PX 441, a CIFM population printout, shows a population figure of only 2176, almost six hundred inmates less, for the relevant day during that period.

Furthermore, Gaes stated that, in light of the number of documents missing from defendants' files on violent incidents, "if, for some reason, the missing reports were somehow related to the actual infraction rate or the population level, then this analysis would be questionable." T. 4910. Plaintiffs' evidence indicates that the rate of missing injury reports has increased substantially from January 1985 through January 1987, PX 418, concurrent with the increase in population.

In sum, based on the evidence, including the testimony of correction officials and officers familiar with CIFM as well as the testimony of Professor Cox and Dr. Gaes, I conclude that overcrowding is a significant cause of violence at CIFM. Although the historical data from CIFM was not conclusive, except during periods of unrest at CIFM the data tended to show an overall correlation between population level and inmate violence, consistent with the impressionistic testimony of those who work there. While I was impressed with Dr. Gaes' obvious intelligence and the sophistication of his research, and appreciated his efforts to "add another voice"[40] to the court's understanding of the effects of overcrowding, I conclude first, that his views differ less significantly from Cox' than defendants argue, and second, that where their opinions do diverge, Cox' years of experience and pioneering work in this field give his testimony more weight. It is, perhaps, worth noting that, although Gaes criticized Cox for being overly theoretical and removed from reality, the practical realities testified to by CIFM inmates, officers and corrections officials tended to support Cox rather than Gaes.

 
B. Lack of Sufficient Cell Space at CIFM

CIFM has 136 single cells to serve its population of around 2500-2600; the cells also serve the 800 sentenced inmates of the Brooklyn Correctional Facility. T. 3564 (Koehler). Plaintiffs' experts termed this dormitory-cell imbalance a substantial cause of violence because of the high social density in dormitories: dormitory housing "provides more opportunity for inmates to ... engage in assaulting behavior." T. 3984 (Bair). For these reasons, the ACA standards for detention and correctional facilities require that inmates in multiple occupancy housing be "screened for suitability to group living prior to admission." PX 372-73. As to detention facilities, the ACA states that dormitory housing should be used for minimum security inmates only. At CIFM, inmates of all security classifications have been housed in dormitories. Defendants' expert Chief Painter also stated that dormitory housing is more conducive to exploitative behavior than cells, even when inmates are classified, T. 4621-22.

Lack of sufficient cell space has serious consequences at CIFM. As discussed in Section I(C) (1) of this decision, plaintiffs' experts Bair and Shoultz were both struck by how inmates with extensive histories of violence were repeatedly returned to general population housing at CIFM, where they committed further acts of violence. T. 3782 (Shoultz); T. 4015-19 (Bair); see also *1547 PX 376. Based on their general correctional experience and on the large amount of violence that has been documented in CIFM's protective custody dormitories, both experts condemned the use of dormitories for protective custody, which neither had ever seen done at other institutions. T. 3808-09 (Shoultz); T. 4145 (Bair). Bair testified that, for CIFM's rated capacity of 2083, between 300-400 cells would be required for inmate safety, based on his general experience of the number of cells required for inmate safety as well as his knowledge of CIFM's history of "entrenched violence." T. 3986-88.[41]

At trial, Commissioner Koehler testified that there were no plans to make available additional cell space for sentenced inmates. T. 3548. He explained that dormitories are "easier to build, much faster to build, and they are less expensive," but added that if he "had [his] choice, [he] would build all cells." T. 3550. In his post-trial supplemental affidavit, Koehler states that "the Department's Central Punitive Segregation Area, which is at the Rikers Island House of Detention for Men ("HDM"), will be expanded from 24 cells to 300 cells. As a result, the number of cells available to house inmates guilty of infractions, including CIFM inmates, will increase." Supp. Aff. at ¶ 16. However, while this measure may alleviate the situation slightly, it is doubtful whether it will have a significant effect, because few violent incidents at CIFM result in a sentence of punitive segregation, and when a punitive sentence is over, the inmate will presumably have to return to a general population dormitory.

I conclude that lack of sufficient cell space and overreliance on dormitory housing is a significant cause of inmate-inmate violence at CIFM.

 
C. Lack of Classification

There is no dispute that lack of classification at CIFM has been a significant cause of violence there. Classification reduces violence because violent inmates are less likely to commit violent acts against others like themselves, and because, once violent inmates are identified and grouped, it is possible to take appropriate control measures such as putting them in single cells and making changes in staffing and programming. T. 3795-96 (Shoultz); T. 3992-93 (Bair). All the corrections experts agreed that lack of classification was a substantial cause of violence at CIFM, both among inmates and staff, T. 3780, 3837-38 (Shoultz); T. 3978, 4026, 4028-29 (Bair); 4576-77 (Painter), and that a good classification system would reduce violence. Bair testified that lack of classification at CIFM was "like playing Russian roulette." T. 3992.

Before the trial, defendants admitted that "[g]eneral population inmates are not classified or separated by security level, length of sentence, crime of conviction, previous criminal record, or previous history of incarceration." PX 247 (admission 48). Instead, inmates were assigned to dormitories according to whether they had jobs and what their job assignments were. The only other distinctions were the separation of adults from adolescents and parole violators from inmates serving city sentences, and the removal of some inmates from general population to "special housing," including dormitories for drug detoxification, mental observation and infirmary use, and to four cell areas, where different groups of inmates, such as those in protective custody and administrative segregation, were kept and sometimes mixed.

Over the past decade, the Department of Correction has attempted at least twice and possibly as many as five times to develop classifications systems at various facilities. T. 2537-43 (Digirolamo). The Department's previous efforts proved to be half-hearted efforts which, as Francesca Digirolamo, the current director of classification *1548 agreed had happened in one instance, "sort of fizzled out." T. 2538. Digirolamo began working on the current classification proposal as part of an effort to comply with various consent decrees applicable to detainees. The record is clear that the extension of this proposal to CIFM was substantially motivated by the litigation in this case. T. 2497 (Digirolamo) (she first began work at CIFM "when [she] was approached with the fact that there was this litigation going on, and that the legal department ... would want [her] to get involved possibly...."); T. 3425 (Koehler) (because of his "understanding that we had this case, understanding that the sentenced inmates were very important, and that we should be doing more classifying of them, I asked [Digirolamo] to divert some of her attention to C-76").

Implementation of the CIFM classification proposal began in May 1987,[42] after a three month period of gathering data. It provides for initial "risk screening" to identify inmates who require special housing, and places general population inmates into low, medium and high security categories, with an additional maximum security special housing category. Classification is assigned through a point system based on various factors, most related to the inmate's criminal record.

Plaintiffs argue that the new classification plan, while considerably better than nothing, has the following flaws:

a) Lack of information about behavior in prior periods of incarceration: Plaintiffs' experts testified that information about behavior during past incarcerations at CIFM would be important given the high rate of recidivism at CIFM, T. 3801-02 (Shoultz); T. 4022-23 (Bair). For instance, inmate Kelly Davenport was involved in twenty violent incidents during an eight month period in 1985; however, when he was reincarcerated at CIFM during the trial in this action, he was placed in a general population dormitory. T. 4021-22 (Bair); PX 376, 384. At the time of trial, information about past incarcerations was not factored into the classification proposal. I note, however, that Commissioner Koehler in his post-trial supplemental affidavit states that he has hired a consultant to "design a component for the existing computer information system to make available information on infractions committed during prior incarcerations" and other types of violent behavior. Koehler Supp. Aff. at ¶ 15b.

b) Lack of criteria for placement in cells versus dormitories: Under the current system, the high, medium and minimum security prisoners may all be housed in dormitories; only maximum security inmates and a few other special housing categories are designated for single cells. Defendants' point system goes from one to thirty-three; only those inmates with a score of twenty-nine or more are classed as maximum security. In the first three months of assigning classification scores, no CIFM inmate scored more than twenty-three, DX DDD at ¶ 10, and hence none of them were classified as maximum security.

c) Lack of a personal interview by classification personnel: Both plaintiffs' and defendants' experts agreed on the importance of a personal interview by qualified classification personnel. T. 3800-02 (Shoultz); T. 4646 (Painter). Defendants argue, however, that although their classification system does not provide for a formal interview, there are other opportunities to communicate with inmates, and "inmates with a particular need will make themselves heard." DX DDD at ¶ 1.

d) Lack of provision for separate housing and placement in cells of special housing categories: Under the current system, inmates with special needs, such as homosexual, protective custody, mental observation, and maximum security inmates, and inmates in various administrative segregation categories, including multiple violent infractors, can all be placed in dormitories *1549 and mixed together. DX HHH ("Guidelines for Special Housing Assignments").

It is too early to determine whether defendants' classification plan is adequate to reduce violence at CIFM significantly. However, plaintiffs' arguments are substantial and will be considered in framing a remedy.

 
D. Inadequate Staffing and Supervision

Plaintiffs' experts stated that inadequate staffing was a significant cause of both staff-inmate and inmate-inmate violence. Most of the evidence presented concerned the link between inadequate staffing and inmate-inmate violence, although plaintiffs' experts did testify that officers who feel overwhelmed and outnumbered are more likely to become frustrated and "trigger-happy" when dealing with inmates. T. 3813, 3839 (Shoultz); T. 3980-81, 4027-28 (Bair).

1. Dormitory Staffing: Most of CIFM's inmates are housed in CIFM's dormitory housing units. Each dormitory consists of a large rectangular sleeping area. At the front of the dormitory there is a dayroom and a bathroom, with a glass-panelled officers' station set further back. Access is controlled by the "A" gate, which leads from the outside corridor into a shorter corridor to the "B" gate near the officer's station. The "B" gate leads into the sleeping area. PX 397 (diagram of CIFM dormitory).

Defendants have used three different staffing patterns at the CIFM dormitories: a single A officer, an A officer and a full-time B officer, and an A officer with a floating B officer who splits his time between two dormitories. The A officer has record-keeping duties and is in charge of inmate movement in and out of the housing area. The B officer, if any, is assigned to remain behind the B gate, that is, in the sleeping area, the dayroom or the bathroom, to monitor the inmates. T. 1073-74 (Officer Simpson). When there is no B officer, the A officer is charged with all of these responsibilities.

Plaintiffs' experts testified that safety requires both an A officer and a B officer at all times, T. 3812-13 (Shoultz); T. 4004-05 (Bair), explaining that it was impossible to observe activity in all parts of a dormitory from the officers' station. On the court's tour, I observed how poor the sight lines were: the presence of bunk beds in the dormitories made visibility especially difficult. A number of injury reports in the record demonstrate the problems caused by lack of staffing. See, e.g., PX 377 (inmate attack occurred at night while lights were out in dormitory, B officer was having a meal, and A officer was unable to see the back of the dormitory); PX 403 (incident 1/87 # 23 (inmate attack occurred at rear of dormitory, out of view of correction officers)).

In 1983-1984, staff cuts led defendants to reduce substantially the officer staffing of the general population dormitories. PX 84. These officers were never replaced. Instead, in December 1986, defendants shifted to a plan of "density driven staffing," under which dormitories with more than seventy inmates receive additional staff: some dormitories share a B officer with another dormitory, and others have been assigned a permanent B officers, depending on the character of the dormitory. T. 3406-08 (Koehler). While this plan improves dormitory staffing, it still leaves many dormitories with only one officer for substantial periods of time.

2. Cell Area Staffing: CIFM has four cell areas: a "V" shaped pair of thirty-four cell corridors on each of two stories of the North Side. Each pair shares one control area or "bridge," where officers are stationed. The dayrooms are at the front of end of the corridors, with large glass panels for observation from the control area. Access to the cell corridors from the control area is through the B gate, and access to the outside corridor is through the A gate. The cell doors are controlled by locks in a panel box set into the wall in the control area outside the B gate. T. 1629-33 (Officer Long).

Until recently, each pair of cell corridors had only one supervising officer each: for *1550 instance, the officer from 2 Main would have to assist the officer from 3 Main when necessary and possible as back-up. As an example of the problems caused at least in part by inadequate cell area staffing, on August 21, 1986, inmate Marco Reyes, while locked in his cell in 2 Upper, had his eye poked out with a broomstick by another inmate who was sweeping the hall, after a verbal dispute. While this was going on, one officer was making a tour of 3 Upper and one was in the control station. PX 69. As another example, on June 14, 1986, two inmates were stabbed in separate incidents in cell area 2 Upper. Neither stabbing was observed by the guards, and one was not even reported until four hours later. In the unusual incident report on one of these incidents, the investigating captain concluded, as did several other officers who submitted reports on the two incidents, that "2 & 3 Upper posts necessitate [another] officer to minimize the recurrence of similar incidents in the future." PX 125B (unusual incident reports 20MM147 and 20MM148).

In April 1986, then-Warden Cowan requested permission to hire officers to act as floating B officers, stating that "[w]ithout these additional posts, inmate safety and supervision cannot be assured." PX 256 at 2 ("Updated New Needs Request"). This request was not granted during Cowan's tenure. Then-Warden Garvey, testifying at trial on December 29, 1986, stated that he too was seeking additional staffing for the cell areas. When asked by the court, "You regard that staffing as really essential, ... for maintaining safety"? He answered "Yes.... We left ourselves vulnerable." T. 2965. As described below, after the conclusion of the trial, some additional staffing was added to the cell areas.

3. Stair and Gate Staffing: Most inmate movement at CIFM is unescorted. On stair landings, a single officer is responsible for traffic in four directions, and no officers are placed in stairwells themselves. T. 3533-35 (Koehler). Many violent incidents occur there: PX 375 lists some eighty-one incidents which took place in the stairwell and gate areas from January 1985 to August 1986, including slashings, stabbings, and attacks by groups of inmates on single victim.

4. Absence of assigned staff from posts: There is evidence of record that housing staff is regularly removed for meals or to perform other duties without relief, a practice which plaintiffs' experts condemned as unsafe. See, e.g., T. 3812-13 (Shoultz).

Defendants argue that numerous incidents of inmate-inmate violence would still occur even if CIFM was more heavily staffed, citing evidence that some assaults take place even in the presence of staff, that some attacks are precalculated and unpreventable, and that some inmates are incorrigible. In addition, Chief Painter testified that the level of staffing at CIFM was adequate and comparable to that in his Los Angeles system, where one officer supervises one or two dormitories with the assistance of roving officers who walk in and out, and some dormitories are left completely unattended for periods of time. Painter also testified that staffing could be reduced when inmates were at work and at night, although he acknowledged that if many assaults take place at night, as they do at CIFM, additional staffing might be necessary. T. 4669-71.

While it is true that some inmate-inmate attacks may be unpreventable regardless of the level of staffing, the evidence indicates serious flaws in CIFM staffing and plaintiffs have convincingly demonstrated a connection between inadequate staffing and the occurrence of inmate-inmate violence.[43] There is no question, however, that since the trial began, the situation has improved. At the beginning of the trial, CIFM had 711 officers; by the end of the trial, there were 796. DX BB; DX XXX at ¶ 4. The new staff includes an additional officer on two of the three daily tours in each of the cell areas and approximately sixty-two new officers assigned to the dormitories, *1551 as well as additional meal-relief/escort officers. Again, these changes will be taken into account in fashioning a remedy in this case.

 
E. Inadequate Systems for Controlling Use of Force

The heart of plaintiffs' case on the causation of excessive force at CIFM is evidence that there have been systemic failures both to take steps to prevent misuse of force and to investigate and discipline those officers who do misuse force. These systemic deficiencies, as discussed below in Section V, provide the basis for a finding of deliberate indifference necessary to state an Eighth Amendment violation for staff-inmate violence. Four problem areas will be discussed: 1) defendants' use of force policy; 2) training of correctional officers on use and misuse of force; 3) investigation and monitoring of misuse of force; and 4) officer discipline for misuse of force.

1. Written Use of Force Policy: Directive 5002, the statement used by the Department of Correction until after the end of the trial in this case to express its policy on staff use of force, provided little else other than that "force of any kind must be used only as a last resort where all other alternatives have been exhausted ...," and that force "is to be minimized as much as possible." The alternatives to force and ways to minimize force were not explored. Plaintiffs' expert on use of force, Vince Nathan, described Directive 5002 as "a vague, generalized statement which provides virtually no guidance to a facility manager, let alone a line correctional officer as to the conduct that's expected," T. 4242, and termed it "the beginning of a predictable phenomenon of excessive force and uncontrolled force through the institution," T. 4251. Defendants' use of force consultant, Michael Gilbert, called it "exceptionally vague." PX 383 at 89. Only Chief Painter argued that Directive 5002 was "pretty specific," stating that greater specificity could be confusing. T. 4574-75.

After the end of the trial, defendants issued a new directive on use of force, Directive 5002R. Koehler Supp. Aff. at ¶ 12 and Exhibit M. This directive is more comprehensive and detailed: it provides definitions and examples of what constitutes use of force, unnecessary use of force and excessive use of force, and it provides examples of alternatives to force. Plaintiffs, however, argue that the new policy, while better than the old, is still not sufficiently detailed, and still provides the "loophole" of allowing an officer to use force "to maintain the good order of the facility," without specification or limitation.[44]

2. Training of Correctional Officers: Plaintiffs argue that inadequate training of recruits is a substantial cause of misuse of force, a contention supported by a report entitled, Excessive Use of Force by Staff: A Training and Management Problem for the New York City Department of Corrections (March 9, 1987), PX 383 at 69-70, written by Michael Gilbert at the request of the Department of Correction. This report concluded that "the current performance problems associate[d] with excessive use of force can be linked, IN PART, to the inadequacy of the formal *1552 training provided to new Correctional Officer employees." Id. (emphasis in original).

At the core of the inadequate training problem is the weakness of the Training Academy. Until after the trial, new correction officers were given an eight week preservice training program. Only three hours of instruction on the justifiable use of force under the state law and departmental policy were taught. T. 3012 (Assistant Deputy Warden Al-Rahman). There were thirty-two hours devoted to self-defense tactics, with some of this time spent on calisthenics. PX 99 at 58-62 (deposition of self-defense trainer). Most of the time was spent in the classroom hearing lectures or watching videotapes, PX 383 at 42, as opposed to participatory activities.

The Training Academy in recent years has begun to train recruits in "Aikido," a self-defense technique that does not rely on punches, kicks or other street-fighting techniques, that is highly recommended both by plaintiffs' expert Bair and by the Gilbert report. PX 383 at 18; T. 4036-37 (Bair); PX 99 at 78-51. However, Bair noted that the training in Aikido seemed to have little effect after officers left the Academy:

 
The reality is what is being used in C76 in almost all occasions, well over 90 percent that I reviewed, was boxing techniques. When the inmates or staff were threatened or assaulted, they would respond by punching the inmates in the face as opposed to blocking and using the [Aikido] technique.

T. 4036. The testimony of several officers indicated that they either were not taught Aikido, had forgotten it, or lacked confidence in their abilities to use it. See, e.g., T. 1674-75 (Officer Long); T. 1242 (Officer Murray). The self-defense trainers deliberately avoided practical, detailed guidance as a matter of policy in order to avoid what one trainer termed "editorializing" to the recruits. T. 3021 (Al-Rahman). Recruits were not even taught that blows with fists or kicks to particular areas of the body should be avoided. T. 3060 (Al-Rahman). In addition, the Academy was severely understaffed, overworking the staff through "an overabundance of overtime," T. 3039 (Al-Rahman), and leading to classes which were too large. For instance, self-defense classes had up to eighty students per session with only two instructors. PX 99 at 30-31, 35-36.

In addition to the problems at the Training Academy, plaintiffs presented evidence that there was insufficient on-the-job training for recruits and in-service training for older officers. Gilbert called the on-the-job training system "a relatively meaningless exercise," PX 383 at 70-71, and recommended a far more comprehensive and structured program. Gilbert also concluded that the lack of in-service training contributed to the excessive force problem by increasing staff frustration, PX 383 at 76-79, 60-61, 73. Commissioner Koehler acknowledged frankly that "[t]he absence of in-service training ... has operated to the detriment of the New York City Department of Correction...." T. 3378.

After the trial, defendants announced changes in the training program, including the following: 1) Commissioner Koehler has received approval to spend $7 million for training in fiscal year 1989, and thirty new instructors have been added to the Training Academy staff, twenty-six of whom are filling new positions; 2) twenty-five experienced officers have been selected and trained to provide on-the-job training through a program whereby recently hired officers will work in housing areas with experienced housing area officers; 3) the new Directive 5002R contains hypothetical scenarios that will be used in both the Academy and in-service training; 4) DOC has hired a consulting firm, "Handle With Care," to train self-defense trainers to train recruits in a special system of self-defense that emphasizes body holds and communication as a first line of defense; 5) the Department of Correction will offer five days of annual training to veteran correction officers on the use of force, non-violent alternatives to force, self-defense techniques, and report writing. Koehler Supplemental *1553 Affidavit at ¶ 14; DX XXX at ¶ 2.[45]

3. Investigation and Monitoring of Use of Force: Plaintiffs' experts Nathan and Bair testified that an adequate system of monitoring, investigation and discipline of misuse of force is essential to keep the use of force within proper bounds. T. 4046 (Bair); T. 4251-52 (Nathan). Although defendants have developed a facially complex system of monitoring, investigating and disciplining misuse of force, the record of this case indicates that it has broken down and functions only minimally.

a) Facility Investigations: Investigations of misuse of force are conducted, at least in the first instance, by the correctional facility in which the incident took place. The system is designed to work roughly this way: First, the use of force is reported to various departmental authorities.[46] Then, the CIFM captain with responsibility on the shift for the area where the incident occurred conducts an investigation, prepares a written report, and attaches supporting documentation. The report is then submitted to the tour commander, who fills out a cover report on a four-part form that is ultimately reviewed and signed by the warden. T. 2946 (Garvey). Despite these complicated and seemingly thorough reporting requirements, however, there is evidence of record of at least a dozen examples, from January to March 1987 alone, where incidents of use of force were not reported. See Plaintiffs' Appendix H ("Selected Examples of Uses of Force Not Reported According to DOC Guidelines").

Beyond the problem of inadequate reporting, the experts for both plaintiffs and defendants agreed that the use of force investigations memorialized in the unusual incident reports were seriously deficient. Nathan and Bair both testified that the reports showed a fundamental bias: "the purpose of the exercise is essentially one of exoneration." T. 4290 (Nathan); T. 4055 (Bair). Bair, who reviews all use of force investigations done within his jurisdiction in Virginia, testified that he would have sent ninety percent of the CIFM investigations back for additional work; during his tenure in Virginia, he has only had to return two. T. 4073-76. Defendants' expert Chief Painter agreed that many of the CIFM investigations were unsatisfactory and said that if he were in charge he would send back "many" if not "most." T. 4635-37.

Some of the specific deficiencies plaintiffs' experts described were: 1) a general lack of thoroughness; 2) failure to pursue unanswered questions; 3) failure to address whether the staff conduct actually complied with policy; 4) inadequate efforts to identify and interview witnesses; 5) failure to explain how conflicting evidence was resolved; 6) a double standard of interpreting medical evidence; 7) magnification of discrepancies in inmate statements and failure to inquire into discrepancies in staff statements; 8) mischaracterization of medical *1554 evidence and of inmate statements; 9) use of boilerplate language; 10) "reaching" for some basis to approve staff conduct; and 11) the conduct of investigations by captains who were directly involved in the incident. T. 4054-73 (Bair); T. 4290-4317 (Nathan).

An example of these problems is the investigation conducted of the incident about which Roger Ramsey testified, where a false alarm was sounded and a number of officers ran into the dormitory past the officer who was trying to tell them it was a false alarm. Ramsey saw them coming and ran to a lower bunk. According to Ramsey and several inmate witnesses, Ramsey was pulled off the bed and beaten. T. 935-43. The investigating captain, DiNapoli, reported that "Officer Vanderpool and two other unknown officers pulled inmate Ramsey ... away from the bed" and concluded that "Officer Vanderpool possibly used force in this situation, in the form of one punch to the face." PX 13. Although the injury to inmate report showed that Ramsay had suffered extensive injuries to his face, back and thigh, and required four sutures for a cut on his lip, PX 14, DiNapoli rejected Ramsey's account that he had been beaten by several officers, stating that "[m]edical findings do[] not reveal him being beaten by several officers." PX 13. The tour commander's report on this incident concludes:

 
Medical evidence indicates that inmate Ramsay did receive injuries on the date in question. It is reported that Officer Vanderpool and two unidentified officers did pull inmate Ramsay from under a bed. Inmate Ramsay could have received injuries noted on "Injury Report" when he was pulled from under the bed. No other force was used according to the reports.

Id. Nathan described the notion that the injuries came from being pulled from under the bed as "a reaching, obviously, on the part of the investigator, to find any conceivable basis to exonerate the officer". T. 4300.

Furthermore, seven inmates signed a statement that Officer Vanderpool and other officers had beaten Ramsay. However, there is no indication that Captain DiNapoli made any effort to interview them and the report does not explain why DiNapoli rejected the statement.[47]

Plaintiffs' experts testified that a large part of the problem is that primary responsibility for the unusual incident investigations rests with the housing captains, who are responsible for the daily supervision of large groups of housing units. First, housing captains have too many other duties. In contrast, when Bair was conducting investigations in Utah, he had no other responsibilities. T. 4053. Also, captains receive almost no formal training in how to conduct investigations. T. 2318-19, 2411 (Captain Cubano). In addition, plaintiffs' experts opined that the objectivity and thoroughness of the housing captains' investigations are compromised because they are investigating the officers whom they supervise and with whom they work every day, and because they are investigating events that may reflect badly on their supervision. T. 4280-82 (Nathan). Chief Painter, however, defended this practice, stating that it is easier for housing captains to spot problems and make adjustments than for outside investigators, and noting that they are immediately available to investigate the incident. T. 4678.[48] The *1555 testimony on this issue indicates that the use of housing captains as investigators may have both positive and negative aspects. From the record of this case, however, it is clear, and indeed substantially undisputed, that the current system of investigations by housing captains is seriously flawed and produces poor results.

b) Investigations Outside the Facility: Until after the end of the trial, the departmental Inspector General was responsible for conducting outside investigations of significant misuse of force allegations at CIFM. All the city departmental Inspectors General reported jointly to the Commissioner of Investigation and to the heads of their respective departments. The correctional Inspector General functioned as part of the Department of Correction, which controlled the office's salaries and promotional opportunities. During the course of the trial, Mayor Koch issued Executive Order 105, which made the Inspectors General responsible solely to the Commissioner of Investigation and limited their responsibilities to allegations of criminal misconduct and official corruption, leaving administrative disciplinary matters to the various city departments. At the time of his testimony at trial, Commissioner Koehler stated that the Inspector General might still retain substantial involvement in investigating misuse of force allegations and testified that he was negotiating with the City with that goal in mind. T. 3371-75, 3488-94.

However, as it turned out, the Department of Correction has assumed responsibility for investigating and overseeing facility investigations of alleged misuse of force. The Inspector General's office will handle only those charges of misuse of force that could lead to charges of criminal misconduct. The Department has created a new unit, the Investigation and Discipline Division, which will combine both investigative and prosecutorial functions for departmental matters, including use of force. The unit will have twenty-one positions for investigators, with a twenty-four hour field command unit on Rikers Island. Although his testimony indicated a preference for more involvement by the Inspector General in use of force investigations, Commissioner Koehler stated in his supplemental affidavit that he believes this arrangement will lead to "efficient and effective employee discipline." Koehler Supp. Aff. at ¶ 6 and ¶¶ 1-10.

The creation of the new departmental division of Investigation and Discipline renders a detailed discussion of the evidence supporting plaintiffs' criticisms of the Inspector General's investigations unnecessary. However, the record on this issue will be briefly described, because it remains relevant to findings of liability and causation, although plaintiffs' claims for injunctive relief against the Inspector General appear to be moot.[49] Plaintiffs' experts noted the following phenomena: 1) in recent years there was a remarkable decline in the number of investigations by the Inspector General, although the number of use of force incidents increased; 2) there was an inadequate flow of information from CIFM to the Inspector General, making selection of appropriate cases to investigate difficult; 3) the Inspector General sometimes failed to follow up even on explicit facility referrals. As to those incidents which were investigated, Nathan, Bair and Painter all agreed that the investigations were of uniformly poor quality, with many of the same problems as the facility unusual incident reports. The reports reflected, among other defects, the investigators' poor training. T. 4339-47 (Nathan); T. 4173 (Bair); T. 4637-38 (Painter). Finally, although drug investigations are completed in thirty days, T. 3231-37 (former Inspector General Judith Schultz), force investigations have not been given *1556 similar priority and, instead, have dragged on for months.

 
4. The Discipline of Staff:

Until recently, responsibility for the prosecution and discipline of officers who are found upon investigation to have misused force has rested with the Advocate of the departmental Inspector General's office. Nathan described the Advocate's office as "ineffective." T. 4353. Now these duties have been transferred to the new Department of Investigation and Discipline, headed by Bonnie Nathan, who had been the Department Advocate.

The discipline system does not appear to have changed in substance despite the transfer in responsibilities. The process starts with the submission to the Commissioner of a "memorandum of complaint" by the warden or another departmental official. After approval by the Commissioner or the Chief of Operations, charges are issued and served upon the charged officer. A conference is then held for discovery and to discuss settlement. Most cases end in pleas. If there is no plea, the case is tried before the city's Office of Administrative Trials and Hearings ("OATH"). OATH officers make findings on merits and recommend a penalty. The Commissioner can overrule either the findings or the penalty, and the process is subject to judicial review in the state courts. Officers can be suspended for up to thirty days pending a disciplinary proceeding. T. 3188-97, 3205-07 (Judith Schultz).

Substantial evidence was put forward at trial that the discipline process is plagued with problems. There are long delays at every stage, including service of the complaint. In many instances, after cases are filed no action is taken until the officer agrees to plead guilty with a light penalty or voluntarily resigns. There are inexplicable delays and sometimes complete failures to act on investigative findings that force was misused. Sometimes departmental charges are never initiated because the officer is given a command discipline, despite the prohibition against this use of lesser sanction in an excessive force case.[50] T. 3306-08 (Judith Schultz); see, e.g., PX 320.

Defendants produced fifty files of disciplinary prosecutions arising out of thirty-six use of force incidents going back to December 1981, which plaintiffs have summarized in Appendix I to their post-trial memorandum. At the end of the trial, twenty-two of these cases were still pending; some arose from incidents occurring more than a year previously and others from as far back as December 1984. By contrast, defendants' expert Chief Painter testified that in the Los Angeles County jail system, disciplinary charges against officers are typically completed within four to six weeks after an investigative finding that force has been misused, T. 4640-41, although defendants point out that there is nothing of record to indicate whether Los Angeles officers are regularly represented by counsel, as are New York correction officers. Of the twenty-eight cases on record which had been disposed of at the time of the trial, only one was the subject of an OATH trial four years earlier. DX TT; Appendix I.

Defendants respond that these delays are often due to factors beyond the Department of Correction's control, such as deferral of investigation while federal or state law enforcement bodies conduct their investigations. In addition, since the conclusion of the trial, the Department has taken steps to speed up the process: officers now appear for informal conferences even before formal charges are served, the disciplinary process has been computerized, and the legal staff has been increased. Since the end of the trial, eleven more misuse of force cases have been resolved or at least closed. Koehler Supp. Aff. at ¶ 8a-e.

Taking an overview, plaintiffs' expert Nathan saw a causal relation between the failures in reporting, monitoring and discipline and the use of force at CIFM:

 
*1557 To the extent that the discovery and punishment of inappropriate conduct is intended to deter conduct of similar type, the deficiencies that I have tried to describe essentially make it clear to an officer, if he wants to notice it, that he will have great latitude in using force and abusing his privilege to use force, that he runs very little likelihood of apprehension, very little likelihood of punishment.

T. 4322. The histories at CIFM of Officers Bryan Bland and Harry Johnson, among others, illustrate how poorly these systems work.

Officer Bland was involved in at least fourteen use of force incidents at CIFM from April 1982 to March 1986. For example, at the trial, inmate Rory Hartley testified that Bland and other officers fractured his cheek, T. 210; Officer Decicco testified that he saw Bland kicking another inmate who was lying on the floor, T. 1861-62; and Officer Torres testified that, on another occasion, he saw Bland jump over a counter and punch an inmate with whom he was arguing, T. 2128-29, breaking the inmate's nose. CIFM's systems for monitoring, investigating and disciplining excessive force failed to curb Bland's proclivities toward violence. Although investigators questioned his judgment on several occasions, Bland received only one forty-five day suspension during the four year period; although discipline or further investigation was recommended in several other instances, these recommendations were never followed up. In October 1986, Bland was transferred to the Queens House of Detention, in response to Commissioner Koehler's request, following the October 1986 disturbances, that wardens suggest officers for transfer who were "viewed as problematic" with regard to the use of force. T. 3420, 3416-20, 3443-44 (Koehler). After the end of the trial, with further charges pending against him, Bland resigned Koehler Supp. Aff. ¶ 8(d).

From July 1982 to October 1986, Officer Johnson was involved in at least twenty-one unusual incidents involving use of force. At trial, inmate Kenny testified that he saw Johnson slam an inmate's head against the wall outside the clinic. T. 263. Inmate Castro said that he was assaulted by Johnson and that he also had witnessed him assault two other inmates. T. 891-94, 897-98. Inmate Crosby testified that Johnson hit him on head with a book. T. 86-87, 89. Although 1) an investigator commented in one of these incidents on Johnson's "negative attitude," 2) Johnson received several corrective interviews for other rules violations in connection with these incidents, and 3) in one instance, departmental charges for unnecessary use of force were specifically recommended, no disciplinary charges were brought against Johnson for use of force. In October 1986, Johnson was transferred to another institution pursuant to Commissioner Koehler's directive that Rikers Island wardens transfer "problematic" officers to other institutions. Nathan termed Johnson's record one of "gratuitous brutality," T. 4270, stating "[t]hat an officer with Mr. Johnson's record could remain employed in an institution or in a department suggests ... that the efforts to control, monitor, investigate, discipline and control the use of force are altogether bankrupt." T. 4275.

Finally, defendants have taken steps during the course of the trial in this action which may help to reduce the incidence of use of force in the future. Commissioner Koehler testified that the Department of Correction has 1) increased the standards by which correction officers applicants are examined and tested, T. 3387-93; 2) increased the probationary period for new officers, T. 3395; 3) implemented an early warning system to assist staff experiencing personal or professional problems that can lead to staff misconduct, T. 3399-3405; 4) taken steps to reduce the amount of overtime work by staff, DX XXX at ¶ 4(d); and 5) instituted a central monitoring system, whereby all correctional facilities will maintain a computer file with entries for each officer who has used force; when any officer uses force on three separate occasions, the warden will interview the officer to ascertain if any problem has developed, T. 3339-3344; Directive 5003. Again, *1558 these changes will be taken into account in framing an injunction.

I conclude that the systematic deficiencies in the four areas described above are a significant cause of misuse of force at CIFM. CIFM's failure to guide and train its officers in the correct use of force and its failure to monitor, investigate and discipline misuse of force have allowedand indeed even made inevitablean unacceptably high rate of misuse of force by staff on inmates.

 
* * *

As a final issue on the subject of causation, defendants argue that the evidence of record indicates that much of the violence at CIFM is ineradicable and irremediable, largely because of the difficulties of managing a population that includes parole violators and adolescents. However, Commissioner Koehler testified that the city-sentenced inmates who compose most of CIFM's population, the majority of whom have been convicted of non-violent crimes and are incarcerated for short but definite periods of time, are easier to manage and control than detainees. T. 3560-62. Even more important, however, it is defendants' duty to protect these inmates adequately unless it is impossible to discharge that duty, and even defendants do not contend that it is impossible. Indeed, the recent changes that defendants have instituted at CIFM show that improvements are practical and possible.

 
V. Legal Standards 
A. Liability

Plaintiffs argue that violence, both inmate-inmate and staff-inmate, pervades CIFM to such an extent that it violates the Eighth Amendment's prohibition against the infliction of "cruel and unusual punishments." The Supreme Court's decision in Rhodes v. Chapman, 452 U.S. 337, 101 S. Ct. 2392, 69 L. Ed. 2d 59 (1981), sets the standard for determining whether plaintiffs have prevailed on their Eighth Amendment claim. In Rhodes, a case involving double-celling in the Ohio prison system, the Supreme Court considered the circumstances under which adverse prison conditions riseor fallto the level of a constitutional violation. The Court began by explaining:

 
The Eighth Amendment, in only three words, imposes the constitutional limitation upon punishments: they cannot be "cruel and unusual." The Court has interpreted these words "in a flexible and dynamic manner," and has extended the Amendment's reach beyond the barbarous physical punishment at issue in the Court's earliest cases. Today the Eighth Amendment prohibits punishments which, although not physically barbarous, "involve the unnecessary and wanton infliction of pain," or are grossly disproportionate to the severity of the crime. Among "unnecessary and wanton" inflictions of pain are those that are "totally without penological justification."
 
No static "test" can exist by which courts determine whether conditions of confinement are cruel and unusual, for the Eighth Amendment "must draw its meaning from the evolving standards of decency that mark the progress of a maturing society." The Court has held, however, that "Eighth Amendment judgments should neither be nor appear to be merely the subjective views" of judges. To be sure, "the Constitution contemplates that in the end [a court's] own judgment will be brought to bear on the question of the acceptability" of a given punishment. But such "judgment[s] should be informed by objective factors to the maximum possible extent."

452 U.S. at 345-46, 101 S. Ct. at 2398-99 (citations omitted).

The Court held that conditions of confinement may constitute cruel and unusual punishment:

 
These principles apply when the conditions of confinement compose the punishment at issue. Conditions must not involve the wanton and unnecessary infliction pain, nor they may be grossly disproportionate to the severity of the crime warranting imprisonment.

452 U.S. at 347, 101 S. Ct. at 2399. Noting that in Estelle v. Gamble, 429 U.S. 97, 97 *1559 S.Ct. 285, 50 L. Ed. 2d 251 (1976), and Hutto v. Finney, 437 U.S. 678, 98 S. Ct. 2565, 57 L. Ed. 2d 522 (1978), Eighth Amendment violations were found where prison conditions caused "pain without any penological purpose," or where they "resulted in unquestioned and serious deprivations of basic human needs," the Court made clear that:

 
[c]onditions other than those in Gamble and Hutto, alone or in combination, may deprive inmates of the minimal civilized measure of life's necessities. Such conditions could be cruel and unusual under the contemporary standard of decency that we recognized in Gamble.... But conditions that cannot be said to be cruel and unusual under contemporary standards are not unconstitutional. To the extent that such conditions are restrictive and even harsh, they are part of the penalty that criminal offenders pay for their offenses against society.

452 U.S. at 348, 101 S. Ct. at 2400.

Finally, Rhodes emphasized that the lower courts, while obligated to enforce the Constitution, must take care not to confuse the requirements of the Constitution with their own personal views as to how jails and prisons should be run. Although "[w]hen conditions of confinement amount to cruel and usual punishment, `federal courts will discharge their duty to protect constitutional rights,'" 452 U.S. at 352, 101 S. Ct. at 2402 (citation omitted),

 
[i]n assessing claims that conditions of confinement are cruel and unusual, courts must bear in mind that their inquiries "spring from constitutional requirements and that judicial answers to them must reflect that fact rather than a court's idea of how best to operate a [correctional] facility."

452 U.S. at 351, 101 S. Ct. at 2401 (citation omitted).

Plaintiffs' Eighth Amendment claims must be examined with these precepts in mind.

 
1. Inmate-Inmate Violence at CIFM:

The Rhodes court included "violence among inmates" as one of several deprivations of essential human needs constituting "conditions intolerable for prison confinement." 452 U.S. at 348, 101 S. Ct. at 2400; see also Inmates of Occoquan v. Barry, 844 F.2d 828, 836 (D.C.Cir.1988). The Court of Appeals for this circuit, in accordance with other circuit courts throughout the country,[51] has recognized that the failure of those charged with the care of prisoners to employ reasonable measures to protect inmates from violence at the hands of other inmates, acting either intentionally, with reckless disregard or with deliberate indifference, constitutes cruel and unusual punishment. See Stubbs v. Dudley, 849 F.2d 83, 85-86 (2d Cir.1988); Villante v. Dept. of Corrections, 786 F.2d 516, 523 (2d Cir.1986); Ayers v. Coughlin, 780 F.2d 205, 209 (2d Cir.1985).

The source of the duty to protect inmates from violence is clear:

 
Having incarcerated the individuals, stripped them of all means of self-protection, and foreclosed access to private aid, the state is constitutionally required to provide prisoners with some protection from the dangers to which they are exposed. Although the state is not obliged to insure an assault-free environment, a prisoner has a constitutional right to be protected from the unreasonable threat of violence from his fellow inmates.

Morgan v. District of Columbia, 824 F.2d 1049, 1057 (D.C.Cir.1987) (citation omitted); see also Doe v. New York City Dept. of Social Services, 649 F.2d 134, 141 (2d Cir. 1981) ("When individuals are placed in custody or under the care of the government, their governmental custodians are sometimes charged with affirmative duties, the nonfeasance of which may violate the constitution. Thus, non-performance of such custodial duties has been held to give rise *1560 to a § 1983 cause of action for prisoners.") (citation omitted).

Although "[t]he cases defining the standard of official liability for inmate-on-inmate violence use slightly different rhetoric [, they] articulate essentially the same standard," Goka v. Bobbitt, 625 F. Supp. 319, 321 (N.D.Ill.1985), which requires that "[t]o obtain relief, an inmate must show `a pervasive risk of harm to inmates from other prisoners,' and that the prison officials have acted with `deliberate indifference' to the danger." Riley v. Jeffes, 777 F.2d 143, 147 (3d Cir.1985) (citations omitted).[52] Accordingly, the question to be decided is whether plaintiffs have demonstrated a "pervasive risk of harm" and "deliberate indifference" on the part of the defendants.

 
a) Pervasive Risk of Harm:

In order to establish an Eighth Amendment violation, plaintiffs must establish that risk of inmate-inmate violence is a "serious problem of substantial dimensions," although violence need not be "rampant." Withers v. Levine, 615 F.2d 158, 161 (4th Cir.), cert. denied, 449 U.S. 849, 101 S. Ct. 136, 66 L. Ed. 2d 59 (1980). Although pervasive risk of harm "may not ordinarily be shown by pointing to a single incident or isolated incidents, ... it may be established by much less than proof of a reign of violence and terror...." Id.; accord Riley v. Jeffes, 777 F.2d 143, 147 (3d Cir.1985); Shrader v. White, 761 F.2d 975, 978 (4th Cir.1985).

Here, plaintiffs have established that inmate-inmate violence at CIFM is a substantial and widespread problem. In 1986, there were roughly 1300 reported violent incidents in an institution with a daily average population of 2500-2600, a figure which does not account for the phenomenon of underreporting because of fear of reprisal which so many courts have recognized. Defendants correctly point out that of the roughly 25,000 inmates who pass through CIFM in a year, the overwhelming majority are not assaulted or injured. Assuming that each item in plaintiffs' table of violent incidents is correct, and assuming that no inmate was involved in more than one fight or assault, on the average from 1982 through March 1987, "only" 1100 out of 25,000 inmates were involved in a fight or assault a year.[53]

However, defendants' comparison of the number of incidents to the total number of inmates who pass through CIFM yearly is inapt: in determining the pervasiveness of violence at CIFM, the relevant consideration is the risk of violence faced each day by each of the 2500-2600 inmates incarcerated, on average, at CIFM on any given day. As plaintiffs put it, "[v]iolence, unlike [other adverse] conditions ..., has an impact that is immediate and not cumulative; an inmate who gets his face slashed is in the same position whether he has been in jail for a day, a month, or a year."[54]

Furthermore, plaintiffs' experts gave reasoned testimony that the level of inmate-inmate violence at CIFM far exceeds that at many of the institutions which they have studied or where they worked: roughly, violence occurs at CIFM at a rate nine times higher than at the Virginia prison system of which Bair is a regional administrator, eight and a half times higher than at Fishkill Correctional Facility in New York, and three and a half times higher than at the Los Angeles jail system. Violence *1561 at CIFM also exceeds that at New York City's borough houses of detention.

Even accounting for differences in reporting systems, these comparisons are both sobering and revealing, and they are legally significant, given the Rhodes court's direction to consider "objective factors to the maximum possible extent" in considering whether prison conditions violate the Eighth Amendment. 452 U.S. at 345 (citation omitted). Several courts have declined to find unconstitutional levels of inmate-inmate violence where there were either no findings of a comparative nature or where those findings indicated only that there were comparable levels of violence at comparable institutions. See, e.g., Inmates of Occoquan v. Barry, 844 F.2d 828, 832 (D.C.Cir.1988); McGriff v. Coughlin, 640 F. Supp. 877, 880 (S.D.N.Y.1986); but cf. Doe v. District of Columbia, 701 F.2d 948, 964-65 (D.C.Cir.1983) (statement of Edwards, J.) (noting that "it is clearly erroneous to assume that average' conditions will invariably pass constitutional muster").[55]

Furthermore, looking beyond statistics, the testimony heard and evidence received at trial show that serious violence has become almost a feature of institutional life at CIFM. Aggressive inmates with a propensity to violence, some of whom gang together in "posses," have not been kept apart from the general population, even after repeated acts of violence. Although there have been no deaths at CIFM during the period examined, and fewer rapes have occurred than described in some other cases where Eighth Amendment violations have been found, inmates have suffered slashings, burnings and other forms of violent injuries with chilling regularity. Only fortuity has saved some of these inmates from death. Many of these injuries were inflicted by weapons, both commercial and homemade. The attacks have taken place not only in general population dormitories, but also in protective custody, where inmates who fear for their personal safety are housed for protection, and in the halls and stairways of the institution, through which all inmates must pass on their way to meals and other activities.

 
b) Deliberate Indifference:

The deliberate indifference standard was first articulated in Estelle v. Gamble, 429 U.S. 97, 104, 97 S. Ct. 285, 291, 50 L. Ed. 2d 251 (1977) (citation omitted), when the Supreme Court concluded that "deliberate indifference to serious medical needs of prisoners constitutes the `unnecessary and wanton infliction of pain' proscribed by the Eighth Amendment." The deliberate indifference standard has since been applied to prison conditions cases involving other essential services necessary to an inmate's well-being, including personal safety. See Benson v. Cady, 761 F.2d 335, 339 (7th Cir.1985).

The Second Circuit's decision in Todaro v. Ward, 565 F.2d 48, 52 (2d Cir.1977), affirming a determination that inadequate medical services violated the Eighth Amendment at a New York State prison, provides guidance in seeking the definition of "deliberate indifference" in a prison conditions case:

 
[W]hile a single instance of medical care denied or delayed, viewed in isolation, may appear to be the product of mere negligence, repeated examples of such treatment bespeak a deliberate indifference by prison authorities to the agony engendered by haphazard and ill-conceived procedures. Indeed, it is well-settled in this circuit that "a series of incidents closely related in time ... may disclose a pattern of conduct amounting to deliberate indifference to the medical needs of prisoners." When systematic deficiencies in staffing, facilities or procedures make unnecessary suffering inevitable, a court will not hesitate to use its injunctive powers. *1562 (citations omitted).[56]

Under Todaro, "systematic deficiencies" at CIFM, of which the defendants have had ample notice and which make "unnecessary suffering inevitable," provide evidence of deliberate indifference on the defendants' part. See also Eng v. Smith, 849 F.2d 80, 82-83 (2d Cir.1988) (affirming preliminary injunction in case alleging Eighth Amendment violations in mental health care at Attica Correctional Facility; deliberate indifference requirement met by showing of "systemic deficiencies" in mental health care system). The evidence summarized above of persistent overcrowding, failure to classify, excessive reliance on dormitory housing and inadequate staffing, in the face of clear warnings that such conditions were exacerbating violence at CIFM, provide sufficient evidence of deliberate indifference on the part of the defendants.

This finding should not be interpreted as casting any cloud on the obvious sincerity and competence of Commissioner Koehler, then-Warden Garvey, and many of the other CIFM officers and officials from whom we heard testimony. A finding of "deliberate indifference" from evidence of systematic deficiencies in prison conditions is not inconsistent with findings that individual officials or staff members at an institution attempt in good faith to do a good job. For instance, the district court in Todaro concluded by noting that it was

 
not unfavorably impressed with the individual members of the ... medical staff. They appeared to be truly concerned with the well-being of the inmates they served. Further, the Court agrees with defendants that the medical care provided at [the prison] is nowhere near as shockingly inadequate as that provided in other institutions described in the cases cited in this opinion ... and that adequate medical care is often provided at [the prison].

Todaro v. Ward, 431 F. Supp. 1129, 1159-1160 (S.D.N.Y.1977), aff'd, 565 F.2d 48 (2d Cir.1977); see also Grubbs v. Bradley, 552 F. Supp. 1052, 1079 (M.D.Tn.1982) (unconstitutionally inadequate protection of inmates caused by "deficiencies [which] appear to be endemic to the system, and to have roots much deeper than the conduct of individual officers"). As Justice Brennan has noted, "even conscientious prison officials are `[c]aught in the middle,' as ... legislatures refuse `to spend sufficient tax dollars to bring conditions ... up to minimally acceptable standards.'" Rhodes v. Chapman, 452 U.S. at 358, 101 S. Ct. at 2405 (Brennan, J., concurring) (citations omitted). So too, the endemic inmate-inmate violence at CIFM is the result of systematic deficiencies and neglect, reflecting an institutional failure on the part of New York City that cannot be wholly attributed to individual officials at CIFM.

 
c) Excessive Force at CIFM:

The Court of Appeals for this circuit, recognizing that prisoners "are at the mercy of their keepers," has found that use of force by staff against inmates that is "wholly beyond any force needed to maintain order" violates the Eighth Amendment. Inmates of Attica Correctional Facility v. Rockefeller, 453 F.2d 12, 23, 22 (2d Cir.1971); see also Hoptowit v. Ray, 682 F.2d 1237, 1251 (9th Cir.1982) ("Physical brutality cannot be part of the daily routine. *1563 Rather, guards may use force only in proportion to the need in each situation.").

In determining whether the line has been crossed between legitimate and excessive force, the court must consider such factors as

 
the need for the application of force, the relationship between the need and the amount of force that was used, the extent of injury inflicted, and whether force was applied in a good faith effort to maintain or restore discipline or maliciously and sadistically for the very purpose of causing harm.

Johnson v. Glick, 481 F.2d 1028, 1033 (2d Cir.), cert. denied, 414 U.S. 1033, 94 S. Ct. 462, 38 L. Ed. 2d 324 (1973).[57]

Applying these guidelines to the testimony heard and evidence received at trial, it is determined that plaintiffs have shown a pattern of misuse of force by officers at CIFM. Although plaintiffs' statistical evidence on staff-inmate violence is more difficult to interpret than the statistics on inmate-inmate violence,[58] the record is replete with incidents, in many cases substantially undisputed by defendants,[59] where the use of force by CIFM correction officers was far in excess to the need for force, with serious injuries resulting. In numerous incidents the officers acted, if not "maliciously and sadistically," with at least a viciousness going beyond "a good faith effort to maintain or restore discipline." Defendants' own expert, Chief Painter, called use of force at CIFM "higher than it might be," T. 4585, and agreed that it was "higher than it needs to be," T. 4631.

It is beyond dispute that inmates at CIFM do not suffer the torture and brutality at the hands of correction officers that was found to exist at the institutions described in some of the seminal Eighth Amendment cases. CIFM is far from being "a dark and evil world completely alien to the free world,'" as the Arkansas prison system was described a decade ago in Hutto v. Finney, 437 U.S. 678, 681, 98 S. Ct. 2565, 2569, 57 L. Ed. 2d 522 (1978) (citation omitted). However, this does not excuse CIFM for allowing the regular use of force against inmates which far exceeds that warranted by the circumstances. Furthermore, the October 1986 disturbances suggest that excessive force at CIFM has the potential to rise to the level of the "barbarous conduct" that occurred after an inmates' riot in 1971 at Attica Correctional Facility and was condemned by the Court of Appeals for this circuit. See Inmates of Attica Correctional Facility v. Rockefeller, 453 F.2d 12, 22 (2d Cir.1971) (concluding that staff-inflicted "beatings, physical abuse, torture, [and] running of gauntlets" which followed 1971 Attica riots "far exceeded what our society will tolerate on the part of officers of the law in custody of defenseless prisoners").

*1564 Plaintiffs have demonstrated not only that the level of use of force at CIFM is excessive, but also that the pervasiveness of staff-inmate violence is the predictable result of defendants' policies and practices, most significantly the defendants' failure to prevent misuse of force by adequately training officers in the appropriate use of force and their failure to deter misuse of force by adequately investigating and disciplining use of force.[60] Evidence of these systematic failures supports a finding of a "policy of deliberate indifference" as to staff-inmate violence on the part of defendants.[61]

During the course of the trial and after the trial, defendants made far-reaching changes in many areas, including training, investigatory procedure and discipline. On the basis of these reforms, defendants argue that "[t]he most striking difference between the conduct of defendants in this case and those in which Eighth Amendment violations have been found are the far-ranging, innovative and comprehensive corrective measures that have been implemented by this Department." Defendants' Post-Trial Memorandum of Law at 394. While these changes, most of which took place during or after the trial in this action, will be given serious weight in determining a remedy, as discussed below, they cannot have the effect of preventing a finding of liability based on prior inadequacies which led to constitutional violations of a recurring nature.

Finally, defendants argue that CIFM is a sanitary institution in good physical condition, whereas in most cases "the courts have considered the issue of force by correction officers as an aspect of the totality of circumstances in a prison," rather than examining the issue in isolation, and "where conditions have been found to violate the Eighth Amendment, excessive use of force by staff has generally been found to be only one of many aspects of the prison's operation which was grossly inhumane." Defendants' Post-Trial Memorandum of Law at 390. However, while the court was impressed by many aspects of CIFM's physical plant during its tour there, the evidence presented at trial concerned only the issue of violence: plaintiffs' claims concerning sanitation, health and other conditions have not yet been tried. Furthermore, while many cases have indeed considered both staff-inmate and inmate-inmate violence within the context of a trial challenging many other prison conditions, it is clear that if one challenged condition is found to violate the Eighth Amendment, it is simply irrelevant that other conditions are not challenged or are found to be constitutionally adequate.[62]

 
* * *

*1565 In sum, plaintiffs have established that both inmate-inmate violence and staff-inmate violence at CIFM have reached proportions which violate the Eighth Amendment. There can be no doubt that contemporary standards of decency are violated when inmates, many of them youths and most of them convicted of non-violent misdemeanors, are frequently subjected to often vicious physical abuse both by other inmates and by correction officers, without regard to their offense against society but simply because they are personally vulnerable or offensive to others in the system.

 
B. Remedy

Having concluded that plaintiffs have established Eighth Amendment violations, it remains to be determined whether an injunction should be entered, and, if so, the scope of that remedy.

 
1. The Parties' Proposals:

Plaintiffs argue that they have established the need for a comprehensive injunction that will "address each element contributing to the violation," Hutto v. Finney, 437 U.S. 678, 687, 98 S. Ct. 2565, 57 L. Ed. 2d 522 (1978). Accordingly, they urge the court to make at this stage detailed findings as to the nature of the necessary remedy as well as to liability. As an example of the detailed nature of their requests, they ask the court to rule at this time that the injunction contain provisions, inter alia, that CIFM must be limited to fifty inmates per dormitory, with a total of 1839 for the whole institution, that double bunking should be prohibited in dormitories, and that CIFM require that CIFM build or annex at least 400 single cells to serve the CIFM population.

Defendants, in contrast, argue that the recent reforms at CIFM, especially in the areas of classification, training, staffing and use of force, render moot plaintiffs' claims as to these areas and make injunctive relief unnecessary. They note that this court has on-going jurisdiction over defendants in this case and argue that the court should defer to the correctional administrators in the Department of Correction.

 
2. Need for an Injunction:

While the recent reforms initiated by Commissioner Koehler are impressive in scope, they do not render moot plaintiffs' claims for injunctive relief. The vast majority of these reforms did not take place and many were not even proposeduntil during or after the trial of this case. It is evident from the record, and in some instances openly admitted, that these reforms were initiated in significant part by the pressures of this litigation.

 
The Supreme Court has made clear that: [i]t is the duty of the courts to beware of efforts to defeat injunctive relief by protestations of repentance and reform, especially when abandonment seems timed to anticipate suit, and there is a probability of resumption.

United States v. Oregon Medical Soc., 343 U.S. 326, 333, 72 S. Ct. 690, 696, 96 L. Ed. 978 (1952). The test for mootness is:

 
a stringent one. Mere voluntary cessation of allegedly illegal conduct does not moot a case; if it did, the courts would be compelled to leave "[t]he defendant ... free to return to his old ways." A case might become moot if subsequent events made it absolutely clear that the allegedly wrongful behavior could not reasonably be expected to recur.

United States v. Phosphate Export Ass'n., 393 U.S. 199, 203, 89 S. Ct. 361, 364, 21 L. Ed. 2d 344 (1968) (citations omitted). See also Allee v. Medrano, 416 U.S. 802, 810-11, 94 S. Ct. 2191, 2198, 40 L. Ed. 2d 566 (1974); United States v. W.T. Grant Co., 345 U.S. 629, 633, 73 S. Ct. 894, 897-98, 97 L. Ed. 1303 (1953).

Although this doctrine developed in the context of antitrust suits brought by the government against private parties, it applies *1566 with equal force in cases such as this. In Jones v. Diamond, 636 F.2d 1364, 1375 (5th Cir.), cert. dismissed, 453 U.S. 950, 102 S. Ct. 27, 69 L. Ed. 2d 1033 (1981), for instance, the Fifth Circuit refused to vacate an injunction in a prison conditions case even though a new jail had been built and occupied. Noting that the jail was built only after five years of litigation, the Court stated that:

 
[c]hanges made by defendants after a suit is filed do not remove the necessity for injunctive relief, for practices may be reinstated as swiftly as they were suspended. The burden is on the defendants to prove that the wrongs of the past could not reasonably be expected to recur.
 
The record of defendants' performance permits us to draw no such conclusion, and the improvements in conditions in February 1979, after five and a half years of litigation, does not eliminate the need for an injunction. The provision of a new and sanitary building does not assure that it will be operated in a constitutional way....

Indeed, just last month the Court of Appeals for this circuit affirmed the granting of a preliminary injunction in a case challenging the mental health services at Attica Correctional Facility, despite defendants' claims of substantial post-trial improvement, because "[a]lthough defendants claim to have voluntarily implemented substantially all of the ordered relief, without a preliminary injunction there is nothing to prevent defendants from abandoning procedures which the court determined to be necessary to protect plaintiffs' constitutional rights." Eng v. Smith, 849 F.2d 80, 83 (2d Cir.1988).[63]

Here, the record as it stands now gives no assurance that the changes in policy and procedure at CIFM have significantly reduced violence at CIFM. Moreover, to the extent the changes have been successful, it can hardly be said that events have "made it absolutely clear," as required by Phosphate Export Ass'n., that inmate-inmate violence and staff-inmate excessive force "could not reasonably be expected to recur." 393 U.S. at 203, 89 S. Ct. at 364. It was clear from his testimony that Commissioner Koehler is attempting in good faith to improve conditions at CIFM, and the determination that many of these efforts have been spurred in part by this litigation is not meant as a criticism: certainly this is a case of better late than never. However, the depressing reality is that while commissioners come and go, problems linger on, and present and future inmates are entitled to the assurance that these problems will be, and remain, redressed. Accordingly, the court is obligated to issue an injunction fitted to the circumstances.

 
3. Scope of the Injunction:

As discussed above, defendants' recent changes in practice neither prevent a finding of liability nor render plaintiffs' claims for injunctive relief moot. They do, however, factor heavily in determining the scope of the injunction which should be entered at this time, especially in light of the warnings and guidance provided by Court of Appeals for this circuit's decision in Dean v. Coughlin, 804 F.2d 207 (2d Cir.1986).

In Dean, the Second Circuit considered the propriety of a preliminary injunction entered after a finding that a New York state prison's dental care system violated the Eighth Amendment. After the district court had entered a general injunction, the parties were given the opportunity to submit a proposed plan for the terms of the injunction, and the state proposed a detailed plan to remedy the constitutional violations found by the court. The district court, however, declined to adopt defendants' *1567 proposed plan and entered an injunction of its own devising.

The Second Circuit vacated the injunction, finding that "the [district] court went too far and too fast in imposing upon the state correctional facility its own ideas of how a prison dental clinic should be organized and administered rather than giving appropriate deference to the State's plan." 804 F.2d at 214. The Dean court made clear that the district court must strike a balance in framing an injunction that will both redress the constitutional violations found and yet accord appropriate deference to the defendants' interests in running their own institution:

 
We have repeatedly been cautioned [by the Supreme Court] (1) not to use a sledgehammer where a more delicate instrument will suffice, (2) not to move too quickly where it appears that the state, in the exercise of its administrative authority, will in its own way adopt reforms bringing its system into compliance with the Constitution, and (3) to give the state a reasonable opportunity to remedy a constitutional deficiency, imposing upon it a court-devised solution only if the state plan proves to be unfeasible or inadequate for the purpose.
 
Restraint and initial deference to state institutional authorities in curing unconstitutional conditions are furthermore advisable as a matter of realism; federal courts lack the facilities or expertise needed for formulation and day-to-day administration of detailed plans designed to assure that a state will provide constitutionally acceptable prison services. The court's limited resources and supervisory facilities must accordingly be restricted to intensive use only in those hardcore cases where a state has failed to come forward with a reasonable plan or remedy. A remedial court order must strike a balance between the court's obligation to identify and take steps toward the elimination of constitutional violations and the state's right to administer its own prisons, with the state continuing to have the discretionary responsibility for devising and carrying out a plan whereby it will operate its facilities in a constitutional manner as directed by the court.

Id. at 213-214 (citations omitted).

Here, Dean counsels against the entry of an injunction adopting the specific proposals suggested by plaintiffs at this time, before defendants have been given an opportunity to submit a reasonable plan for the court's consideration. The recent plans for change at CIFM, even though motivated in part by this litigation, evidence strong leadership committed to reform, giving rise to a hope that New York City's correction system has matured in its ability to run its facilities according to constitutional criteria. It is not unrealistic to predict that this may prove to be a case "where it appears that the [city], in the exercise of its administrative authority, will in its own way adopt reforms bringing its system into compliance with the Constitution." 804 F.2d at 213. If defendants, as in Dean, submit a "carefully and conscientiously formulated remedial plan," id. at 215, such a plan would certainly be accorded the "initial deference" advised by Dean, "with such modifications [by the court] as might appear essential to assure compliance with minimal constitutional requirements." Id.

Dean, however, does not stand for the proposition that courts are to offer defendants carte blanche in framing an injunction. Nothing in Dean casts doubt on the district court's equitable power and duty, once constitutional violations have been found, to implement an injunction that is tailored "to fit the nature and extent of the violation," United States v. Yonkers, 837 F.2d 1181, 1235 (2d Cir.1987), and that promises "realistically to work now." Milliken v. Bradley, 433 U.S. 267, 280 n. 15, 97 S. Ct. 2749, 2757 n. 15, 53 L. Ed. 2d 745 (1977) (emphasis in the original) (citation omitted). Indeed, Dean recognizes the propriety of a "court-devised" injunction where "the state plan proves to be unfeasible or inadequate," 804 F.2d at 213, and this surely encompasses any plan which on its face can reasonably be predicted to be unfeasible or inadequate.

*1568 In this regard it is noted that, while the recent changes at CIFM appear designed to ameliorate three of the major causes of violence at CIFM: lack of classification, inadequate staffing, and inadequate monitoring, investigation and discipline of use of force, defendants still do not appear to have substantially addressed two factors which this opinion finds to be significant causes of violence at CIFM: overcrowding and lack of sufficient cell housing. Defendants are urged to submit a plan which addresses all five factors, and any proposed plan which does not address these latter factors will be scrutinized with a heightened degree of care to ensure its adequacy.

 
* * *

The management of this case, including the pre-trial proceedings, the trial and the post-trial briefs, has been an exemplar of high standards of professionalism and civic responsibility on both sides. The determination that CIFM falls short of constitutional standards does not derogate the strongly held belief that those presently charged with responsibility for its management, and their lawyers, are making extraordinary efforts to meet those standards. At the same time, plaintiffs' counsel have made a significant contribution in marshalling the mass of facts necessary to deal with such a case as this, in making lucid presentation at all times, and in presenting the City itself and the court with a true picture of the conditions at CIFM. With this background, the parties and the court can fashion a remedy which will be just to the plaintiffs without being unreasonably burdensome to the defendants.

This memorandum constitutes the findings of fact and conclusions of law required by Fed.R.Civ.P. 52. The parties are directed to prepare for a conference to determine the contents of an order consistent with this opinion.

IT IS SO ORDERED.

 
APPENDIX

This Appendix consists of various tables compiled by the plaintiffs, which either summarize or illustrate various aspects of the evidence presented in this case. These tables, as discussed in the text of this opinion, are found to be substantially accurate, with certain qualifications noted in the text. These tables are appended to this decision because they are helpful to an understanding of conditions at CIFM.

 
                   TABLE 1
           SERIOUS INJURIES RESULTING
             FROM VIOLENT INCIDENTS
               BETWEEN INMATES
                  August 1986
19 slashes, lacerations or stab wounds requiring
   sutures or emergency room
   treatment
 3 other stab or puncture wound
 1 lost eye
 1 perforated eardrum
 1 perforated eardrum with traumatic
   amputation of part of ear
 1 cracked tooth
 1 other transported to emergency room
   or hospital
             February 1987
16 slashes, lacerations or stab wounds requiring
   sutures or emergency room
   treatment
 2 other stab or puncture wounds
 1 fractured rib, collapsed lung
 1 fractured jaw
 4 others transported to emergency room
   or hospital
 1 knee sprain
 1 neck sprain
 1 loss of consciousness
 1 broken tooth
 1 lost tooth
               March 1987
16 slashes, lacerations or stab wounds requiring
   sutures or emergency room
   treatment
 1 other stab or puncture wounds
 1 perforated eardrum
 1 rib fracture
 
*1569
 1 dislocated shoulder
 1 severe soft tissue injury of eye with
   blurred vision
 1 fractured tooth
 1 front teeth knocked out
 2 others transported to emergency room
   or hospital
Sources: PX 105 (injury reports for August 1986)
         PX 403 (injury reports for February and March 1987)
 
                             TABLE 2
              INCIDENTS OF INMATE-INMATE VIOLENCE
                   INVOLVING USE OF WEAPONS
    August 1986                                  March 1987
Date    Weapon/Injury                  Date   Weapon/Injury
8/1   Slash wounds                     3/1   Padlocks
8/3   Chair                            3/1   Broom handle
8/3   Metal bar                        3/1   Lock in sock
8/4   Chair                            3/2   Chair
8/4   Inmate burned                    3/2   Pen
8/5   Broom                            3/4   Mop Wringer
8/6   Lock in sock                     3/4   Bar
8/8   Inmate stabbed                   3/4   Razor blade
8/8   Razor                            3/6   Broom
8/11  Inmate burned                    3/6   Bleaching solution
8/11  Inmate cut                       3/6   Sharp object
8/11  Urine                            3/10  Mop wringer handle
8/13  Metal door trim                  3/11  Lock
8/16  Inmate cut                       3/12  Radio
8/16  Hard object                      3/13  Hard object
8/17  Stab wound                       3/13  Broom handle
8/18  Pipe                             3/13  Broken stick; lock?
8/19  Broom, tray                      3/13  Chair leg
8/20  Sharp pick                       3/14  Razor, chair
8/20  Hot water                        3/15  Tray
8/21  Razor                            3/15  Razor
8/21  Metal pipe                       3/18  Broomstick
8/21  Broom handle                     3/18  Broomstick
8/21  Broomstick, chair                3/18  Puncture wounds, lacerations
8/21  Razor blade                      3/19  Chair, mop stick
8/22  Metal object                     3/19  Razor
8/22  Stab wounds                      3/20  Broomstick
8/23  Razor                            3/21  Stab wound
8/25  Chair                            3/21  Unspecified weapons
8/26  Unknown object                   3/25  Pen
8/28  Cuts                             3/26  Nail or paper clip
8/28  Stab wounds                      3/28  Sharp object
8/30  Stab wounds                      3/29  Book or newspaper
8/30  Chair                            3/29  Urine and feces
8/30  Razor blade                      3/29  Melted plastic or lit toilet paper
                                       3/29  Same as above
                                       3/30  Blunt weapon
Sources: PX 105 (injury reports for August 1986)
         PX 403 (injury reports for February and March 1987)
 
*1570
                                     TABLE 3
                      Reported Inmate-Inmate Violence at CIFM
                                     1982-1987
                       1982       1983       1984       1985      1986      1987*
Violent Incidents       896        911        949       1155      1318       349
Weapons**               403        375        388        479       656       205
Data taken from PX 101-04 and 403.
                       1982       1983       1984       1985      1986      1987*
1986 figures are totalled from PX 101, 104 and 403.
* 1987 figure is for January 1-March 31.
** Includes violent incidents where weapons were used and infractions for possession of weapons.
 
                                     TABLE 4
                           UNUSUAL INCIDENT REPORTS AND
                               USE OF FORCE REPORTS
                                     1982-87
                     Unusual Incidents*        Uses of           Total Incidents
                                              Force**
1982                     159                     0                    159
1983                     195                     1                    196
1984                     129                     7                    136
1985                     194                    26                    220
1986                     169                    32                    201
1987***                      26                    16                     42
* Source: PX 144-207
** Source: PX 261, 261A
*** January-March only
 
                                  TABLE 5
                      REPORTED INMATE-OFFICER VIOLENCE
                 FROM ALL DOCUMENTARY SOURCES FIRST QUARTER
                                1986 AND 1987
                  Unusual Incidents*        Use of       Total UIR/UOF   Injury Reports***
                                           Force**
Jan. 1986               19                    0                19       24
Feb. 1986               16                    0                16       33
Mar. 1986               21                    2                23       27
   Total                56                    2                58       84
Jan. 1987               11                    2                13       28
Feb. 1987               10                    7                17       36
Mar. 1987                5                    7                12       28
    Total               26                   16                42       92
* Source: PX 144-46, 157-59
** Source: PX 261, 261A
*** Source: PX 404, 405
NOTES

[1] The certified class is defined as all inmates in city custody "who are or will be confined in or receive services from the New York City Correctional Institution for Men, and any of its annexes including the so-called `outer dorms,' outer buildings, or `North Facility'...." A sub-class of detainees was also certified. Stipulated Order Granting Class Certification (June 16, 1983). The outer dormitories are no longer part of CIFM, and there are no longer any detainees at CIFM.

[2] CIFM is also known as and referred to by its departmental designation, C-76.

[3] This determination is made as of the conclusion of the trial in June 1987, when the record in this case was substantially completed. With the exception of several submissions in October 1987 as to changes in policy and practice at CIFM, no evidence has been received regarding current conditions at CIFM.

[4] Notice of Motion, October 3, 1986.

[5] Bair, who is currently responsible for the supervision of the nine wardens who operate nine institutions in his region of Virginia, also worked for six years in the Utah state prison administration, rising to the rank of deputy warden. During his years in the Utah prison system, Bair's duties included working as an investigator of inmate criminal activity, Transcript ("T.") 3956-57, and supervising the department which trained correctional officers, T. 3964. Bair moved to Virginia as a warden in 1984 and was promoted to regional administrator in 1986. Bair testified that he has used force to restrain inmates on several occasions and has supervised other corrections officers in incidents involving use of force. He has also been physically assaulted by inmates at least twice. T. 3966-67.

Colonel Shoultz spent thirty-one years in the United States Army, twenty-six of them as an officer in the Military Police Corps. During the course of his military career, Shoultz served as Provost Marshal of the United States 7th Army in Europe, Chief of Operations for the Military Police in Korea, and supervised numerous military prisons, including Leavenworth Prison. In 1972, Shoultz retired from the military and spent the next nine years as Director of Corrections for the Orange County, Florida jails, where both pre-trial detainees, parole violators, and county-sentenced inmates are housed. He then served as Director of Corrections for the Brevard County, Florida jails.

Both Bair and Shoultz toured CIFM for several days, visiting many of the living areas, including dormitories, cell areas, and corridors, as well as the infirmary, gymnasium, outside recreation area and various administrative offices. T. 3968 (Bair). They also examined large numbers of documents about CIFM given to them by plaintiffs.

[6] Nathan, who was a professor at the University of Toledo College of Law for sixteen years, has served as a consultant for the National Institute of Corrections, the New Mexico Department of Corrections, and several county departments of corrections throughout the country, often as an expert on achieving compliance with consent decrees or internal regulations. He has served as a special master in several cases involving staff use of excessive force, most notably the Ruiz v. Estelle litigation involving the Texas state prison system, and Guthrie v. Evans, involving the Georgia prison system. Nathan spent part of a day touring CIFM, and spent part of a day in the office of the Inspector General, T. 4235-39, and he reviewed, among other documents, numerous CIFM policy directives, use of force logs and records, files indicating prosecutorial actions taken by the Inspector General, and injury to inmate reports. T. 4234-35. Nathan has the reputation as one of the most knowledgeable experts in this field.

[7] Professor Cox, Ph.D. in psychology, trained in basic experimental psychological research and clinical practice. He has published over fifty scholarly papers, including twelve on the effects of crowding, including prison crowding. He has conducted crowding research in at least six federal prisons and several state prisons, and has visited many others. Cox has testified as an expert witness in several prison condition cases, including testifying for the United States Department of Justice in the Ruiz v. Estelle trial. Cox spent a day touring CIFM, T. 4691, and reviewed documentation about CIFM. Cox has no experience or qualifications in administering prisons. T. 4690.

[8] Both Commissioner Koehler and former Warden Garvey have had distinguished careers dedicated to public service. Koehler, who was appointed Commissioner of the New York City Department of Correction in September 1986, received a Bachelor of Arts from John Jay College, a Masters of Science from Hunter College and a law degree from Fordham University. He joined the New York City Police Department in 1967 as a police officer, rising through the ranks to Assistant Chief and participating in the development of selection criteria for officer promotions, negotiations of labor contracts, and an evaluation and improvement of the "911" emergency system. Koehler also served as director of, among others, the Communications Division and the Personnel Department of the Police Department.

Garvey, who was appointed Warden at CIFM in October 1986, received a Bachelor of Arts from John Jay College; at the time of his testimony he was close to completing the necessary credits for a Masters Degree in public administration from John Jay. Garvey entered the Department of Correction in 1964 as a correction officer at CIFM, where he remained for eight years. He was then assigned to Rikers Island Headquarters Command, where he was promoted to captain. In 1977, Garvey was promoted to Assistant Deputy Warden and assigned to the House of Detention for Men ("HDM"). He then moved to the Rikers Island Security Division as the executive officer. From 1981 to January 1985, he ran the Department of Correction Training Academy, and he was Warden of the Correctional Institution for Women from January 1985 to March 1986. On January 4, 1988, Garvey was promoted to Supervising Warden, and Bruce Sullivan became Warden of CIFM, the sixth person to serve as warden there since 1982.

[9] The Custody Division, which runs the Los Angeles County jail system, is one of the eight divisions that make up the Los Angeles County Sheriff's Department. Chief Painter has served in different capacities in the Los Angeles County Sheriff's Department since 1957. From November 1957 to September 1958 he worked as the equivalent of a New York City correction officer in a maximum security facility in Los Angeles. Then, after working briefly in the inmate transportation unit, he became a patrol officer, the equivalent of a New York City police officer. He rose through the ranks in the Patrol Division, and also served in the Detective and Technical Services Divisions. From 1975-1976, he supervised the Training Academy where all new hires, both in the area of corrections and law enforcement, are trained. He was promoted to Chief of the Custody Division in 1983.

[10] Dr. Gaes received his Ph.D. in social psychology in 1981. As a research analyst with the United States Bureau of Prisons, he conducts research concerning program evaluation and other issues pertaining to corrections, including crowding issues. Prior to his current position with the Bureau of Prisons, he was a research analyst at the Otisville federal prison. He has been an expert witness in two cases involving crowding issues at prisons and is the author of fourteen published papers, four of which concern crowding.

[11] A recent study of eighty-five percent of the city-sentenced inmates at CIFM for the months of February, March and April 1987 showed the following rough profile: seventy-one percent were sentenced on misdemeanors, nineteen percent were sentenced on felonies, and ten percent were sentenced on violations. Only about seventeen percent of the city sentenced inmates who were admitted to CIFM from February to April 1987 were serving their first conviction, forty-eight percent of them had prior misdemeanor convictions, twenty-six percent had prior convictions for felonies, and nine percent had prior convictions for violations. Roughly twenty-five percent of the newly admitted inmates had prior convictions for violent offenses. Plaintiffs' Exhibits ("PX") 96-98 (Classification Pilot Statistical Reports).

[12] The housing units are designated by a number and the floor on which they are found, e.g., Upper, Main, Lower.

[13] To summarize each incident of inmate-inmate violence documented in the record of this case would be an impossibly lengthy process. Accordingly, this opinion describes only some of the incidents of violence that were inflicted on six of the twelve inmates who testified. There was no indication at trial that the experiences of these six inmates at CIFM were in any way atypical. The incidents were selected as representative of the pattern of inmate-inmate violence at CIFM. It should be noted that many of the incidents about which testimony was heard were less serious than these; on the other hand, there is also evidence of record of incidents more shocking and grave.

[14] See the appendix to this opinion ("Appendix"), Table 1 (tabulating serious injuries resulting from violent incidents between inmates in three sample months).

[15] See Plaintiffs' Post-Trial Memorandum of Law ("Plaintiffs' Appendix"), Appendix B ("Inmates Burned by Other Inmates").

[16] See, e.g., PX 401 (incidents 3/87 # 22 (inmate raped by inmate with razor); 9/86 # 111 (inmate sodomized at knifepoint)); PX 101 (incident 3/86 # 51 (inmate raped by inmates with razors)); PX 115 (infraction log entries 7063-69 (convictions for gang rape)); PX 402 (incidents 12/86 # 3 (sexual assault); 9/86 # 82 (forced oral and anal sex; gang attack)); PX 409 (unusual incident report 23MM60 (sexual assault)).

[17] See Appendix, Table 2 (tabulating incidents of inmate-inmate violence involving use of weapons in two sample months). Defendants argue that, on closer examination, several of the incidents listed in Table 2 do not actually involve the use of a weapon or even the purposeful infliction of harm. Some of defendants' objections appear to be well-founded. However, even defendants do not dispute that the vast majority of the incidents listed in Table 2 are what they appear to be, and Table 2, even with the questionable items removed, remains a sobering testament to the prevalence of attacks involving weapons at CIFM.

[18] See Appendix, Table 3 (tabulating violent incidents where weapons were used and infractions for possession of weapons, 1982-1987).

[19] See, e.g., PX 106 (injury report 2/86 # 39 (inmate Velez forced to strip nude and then burned on legs and face by four other inmates, including Corey)); PX 138 (inmate Houlder placed in protective custody after threats and extortion by Corey and other inmates); PX 135 (inmate Parker moved to protective custody after being threatened by Corey "and his gang of friends").

[20] Injury to inmate reports are completed whenever an inmate is sent to the clinic because of an injury; CIFM policy is to send all inmates involved in fights to the clinic regardless whether they claim injury. The injury report log lists all inmates sent to the clinic and in some cases briefly states the cause of injury. The infraction log lists each infraction chronologically and includes, inter alia, the inmate's name, the nature of the infraction, and the disposition.

[21] See Appendix, Table 3 ("Reported Inmate-Inmate Violence at CIFM 1982-1987").

[22] Defendants note that CIFM has a more comprehensive system of reporting violent incidents than do some of the other correctional institutions to which CIFM was compared at trial, in the sense that a wider range of incidents are considered within the definition of reportable violence. They argue that, as a result, the number of undocumented incidents of violence at CIFM is small. However, while CIFM's comprehensive reporting system no doubt accounts in part for the high volume of reported incidents, the problem of lost records at CIFM and inmate fear of reprisal still indicate that many violent incidents at CIFM go unreported.

[23] See Defendants' Table 2 ("Severity of Injuries from Inmate-Inmate Incidents"), Defendants' Post-Trial Memorandum of Law at 278.

[24] Plaintiffs have pointed to numerous exhibits where incidents of violence, characterized by either inmates or staff as "horseplay," have involved the use of dangerous weapons or resulted in serious injury. For instance, in PX 409, unusual incident report 23MM65 states that inmate Edward Ramirez required fifty sutures for slash wounds in his back and two other inmates were injured in an incident that the investigating captain concluded started "over horse playing and water throwing."

[25] In addition to the evidence of record in this case of inmate fear of reprisals, numerous other courts have recognized that significant amounts of jail violence go unreported. See, e.g., Alberti v. Klevenhagen, 790 F.2d 1220, 1226 (5th Cir. 1986); McMurry v. Phelps, 533 F. Supp. 742, 753 (W.D.La.1982); Ramos v. Lamm, 485 F. Supp. 122, 141 (D.Col.1979), affirmed in relevant part and vacated in part, 639 F.2d 559 (10th Cir. 1980), cert. denied, 450 U.S. 1041, 101 S. Ct. 1759, 68 L. Ed. 2d 239 (1981); Pugh v. Locke, 406 F. Supp. 318, 325 (M.D.Al.1976).

[26] Bair stated that, in the Virginia reporting system, a "fight" is defined as an incident where two inmates exchange blows, while an "assault" is an incident where one inmate inflicts blows on another, who does not retaliate, or any incident where a weapon is used. T. 3971, 3973-74.

[27] See T. 4739-4741; PX 428-30; DX GGG.

[28] As Fed.R.Evid. 609 recognizes, not all convictions can be used to impeach the credibility of a witness.

[29] Defendants' unusual incident report, DX 90, states that "[i]t cannot be ascertained at this time how and at what time each inmate sustained their injuries." While it identifies Vasquez as a "participant" in the mess hall disturbance, he is not the subject of any specific allegation of misconduct.

[30] The following represent just a few of the many incidents on the record which fall into these four rough categories: 1) Officers Mark Vanderpool and Louis Brignardello opened an inmate's locked cell and beat him despite an order from a superior officer not to open the cell. Vanderpool explained that "No one throw[] water on me and get[] away with it," PX 146 (unusual incident report 23MM7). 2) Several response team officers arrived in a cell area and ordered an inmate to finish an unauthorized telephone call. When he refused, the officers struck him with batons. PX 149 (unusual incident report 21MM146). 3) An officer hit an inmate in the face after the inmate made "a threatening motion with his hands." Then-Warden Cowan sent back the report for further investigation because "a fighting stance does not call for the use of force." PX 159 (unusual incident report 19MM93). 4) The Roger Ramsey incident where Ramsey was beaten after a false alarm is one example of the incidents of record where the arrival of the response team led to excessive force.

[31] See Appendix, Table 4 (Unusual Incident Reports and Use of Force Reports 1982-87).

[32] See Appendix, Table 5 (Reported Inmate-Officer Violence From All Documentary Sources First Quarter 1986 and 1987).

[33] See Plaintiffs' Appendix F ("Investigative Findings of Misuse of Force By CIFM Officers").

[34] Clearly, however, the nature of the force used is different, because this court has not heard testimony about regular, sadistic near-torture, which has been found to have been inflicted by staff on inmates in the Texas system. See Ruiz v. Estelle, 503 F. Supp. 1265, 1302 (S.D.Tx.1980), aff'd in part and rev'd in part on other grounds, 679 F.2d 1115 (5th Cir.1982).

[35] Evidence was also presented at trial that idleness and failure to protect inmate property are causes of violence at CIFM. As to idleness, plaintiffs' experts testified that they observed large numbers of idle inmates at CIFM and heard complaints about lack of jobs and other activities. They testified that idleness contributes to violence by providing more opportunities for confrontations. T. 3823-24 (Shoultz); T. 3981, 3991 (Bair). DX 282a indicates that in April 1987, all 1734 inmates who were deemed "eligible to work" were employed, but the payroll for that period includes no more than about 1300 inmates. It is stipulated that there is no document in evidence reflecting the number of inmates who worked during a particular pay period. DX AAAAA. As to property, plaintiffs argue that if defendants provided inmates with lockers with built-in combination locks, inmate-inmate violence arising from property disputes would be reduced. Plaintiffs also argue that there is no support for the testimony presented by defendants at trial that one cause of inmate-inmate violence at CIFM is the possession of jewelry, sneakers and shoes by inmates, and that removing these items would help reduce violence. T. 1913-14, 2032 (DeCicco); 4582-83, 4588, 4667-68 (Painter). Plaintiffs argue that this position is not supported by the record, because, of the incidents for which injury reports reflect information about the subject matter, the majority did not involve either jewelry or sneakers. See Plaintiffs' Appendix D (subjects of inmate-inmate violence, March 1986-1987). Furthermore, there was testimony that other jails allow inmates to possess these items without serious adverse consequences. See, e.g., T. 4500-01, 4582-83 (Painter) (inmates wear own sneakers and shoes in Los Angeles County). While the record on these issues is not insubstantial, it is not as fully developed as the record on the five problem areas described above. Accordingly, I reach no conclusion on the extent to which idleness and property issues cause violence at CIFM, except to note that these two issues are certainly not as critically significant as the five problem areas discussed above.

[36] Significant to the issue of the link between violence and overcrowding at CIFM specifically, the April 1987 report of the New York State Commission of Correction, stating the results of its investigation of the October 1986 disturbances, concluded that "[t]hese events were precipitated, in part, by inmate overcrowding...." PX 95 at xiv. In addition, the May 1988 Final Report of The Special Committee on Use of Force ("Use of Force Report"), noting that "[a] considerable body of literature as well as the experience of practitioners" indicates that under overcrowded circumstances "the potential for violence increases," at 5, recommended that inmates in dormitories be housed according to the ACA standard of fifty square feet per inmate and fifty inmates per dormitory. Recommendation # 2. (After the October 1986 disturbances, Commissioner Koehler appointed a special committee of concerned citizens with expertise in a variety of fields, to evaluate Department of Correction practices and make recommendations as to how to reduce the occurrence of violent incidents, both staff-inmate and inmate-inmate. The final report of this committee was issued in May 1988. This report will be referred to from time to time where relevant to the issues and evidence in this case, although, after consultation with the parties, it was not made part of the record and is not relied upon as evidence.)

[37] As described earlier, Verne Cox is chair of the Department of Psychology of the University of Texas at Arlington. Trained in both experimental and clinical psychology, he has published over fifty papers, including twelve on crowding which concerned in whole or part the issue of prison crowding. He has testified as an expert witness on behalf of the United States Department of Justice in the Ruiz v. Estelle litigation concerning the Texas state prison system and as a witness for the plaintiffs in the Alston v. Coughlin litigation concerning the New York state Fishkill Correctional Facility.

[38] As described earlier, Dr. Gaes is a Senior Research Analyst at the United States Bureau of Prisons in Washington D.C. He conducts research concerning program evaluation and other substantive issues pertaining to the federal prison system and corrections in general. Prior to this, Gaes was a research analyst at the Otisville federal prison. He has been an expert witness in two cases on crowding, and is the author of fourteen published papers, four of which concern prison crowding.

[39] Dr. Gaes also stated that he had seen other studies indicating that his results "may be restricted to the Federal Bureau of Prisons." T. 4857. However, one of these studies involved a prison that Gaes conceded was at or under capacity during most of the study period. T. 4947-48. The other study, PX 417, by Dr. Chris Innes of the federal Bureau of Justice Statistics, concludes that prison deaths from illness, suicide, homicide, inmate assaults, and "disturbances," have no relation to population density, categorized as "lowest," "low," "moderate," and "high". However, these categories were broadly defined, and did not distinguish between single cells, large dormitories, or other types of housing. Gaes conceded that the study's measure of density "is not representative" of social density, T. 4976, that its definition of spatial density was "not a great definition," T. 4970, and that it confounded the two concepts, T. 4971. Cox stated that the study used "almost a nonsensical criterion ... to define density." T. 4723. In addition, the Innes study 1) relies on responses to a mailed Census Bureau questionnaire rather than independent collection and verification of the data and 2) overutilizes "aggregate data," that is, the sum or average of many individual pieces of data, e.g., average daily population or total assaults for some time period, without controlling for other variables, a methodology which Gaes criticized elsewhere in his testimony. The Innes study aggregated data over a year's time, and the only control variable was for security level.

[40] Gaes stated that he agreed to testify as an expert for the defendants, despite the fact that he could not receive payment from them because he is a federal employee, because he felt that "another voice has to be heard in this process other than the ones that have been heard so far." T. 5022.

[41] The May 1988 Use of Force Report noted that "[i]n addition to the custodial problems dormitory housing creates, an overreliance on dormitory housing limits the Department's ability to isolate inmates who are in need of protective custody or who are a physical threat to other inmates or Correction officers," Finding 1, and recommends that "the ratio of single cells to dormitory beds should reflect inmate classification information and should be properly balanced to ensure the safety of inmates and Correction Officers." Recommendation # 1.

[42] The parties stipulated that, if Digirolamo had been recalled as a witness, she would have testified that the CIFM classification system would have begun in October 1986 rather than February 1987 but the pilot program was delayed at the Legal Aid Society's request. DX DDD at ¶ 17. Plaintiffs deny this charge.

[43] Although, as indicated above, there was less evidence and testimony concerning the link between inadequate staffing and staff-inmate violence, there was sufficient evidence to support the logical connection between the two.

[44] The May 1988 Use of Force Committee Report concluded that:

Despite the substantive improvements in the Department's new use of force policy, its structure and organization diminish its ability to effectively communicate behavioral expectations. Clearly written, easily understood behavioral guidelines, that specify alternatives to the use of force and the circumstances under which Correction Officer use of force is appropriate and inappropriate, are critical if policy is to guide Correction Officer behavior in highly unpredictable and emotionally charged situations.

Finding 5. The Report also notes that:

The Correction Officers receive a mixed message they are expected to control the inmates, but they are not given clear guidance as to how to effect the control. The inmate population also receives a mixed message the [Department] does not condone an unjustifiable or excessive use of force, but the policy does not appear to them to effect the day-to-day operations of the [Department]. The public is poorly served when a policy that involves human rights and public safety fails to achieve both the appearance and the reality of justice and accountability.

P. 36.

[45] The May 1988 Use of Force Report concluded that:

the seven week Academy training period for recruits is not sufficient to adequately convey the values, interpersonal skills, and knowledge required to prepare them for the service delivery demands of the job. The overall perception of the recruit training program among experienced officers is that it focuses heavily on the custodial components of the job and is not relevant to the multi-faceted demands that more accurately define the work of a Correction officer.

P. 28. The Committee also recommended a more extensive in-service training program which "should be based on formal training needs assessments and feedback from annual evaluations based on Department performance criteria that should be established for each rank." Recommendation # 19. Koehler, in his response to the Committee dated May 18, 1988, states that he is now implementing a ten week training program, incorporating three weeks of field training.

[46] Until after the conclusion of trial, only uses of force "which result[] in a medically detectable injury or any degree of physical injury, but ... not necessarily a serious injury" were reported as unusual incidents. PX 236 (Directive 5000 at 2). A use of force report was filed whenever force was used, but these reports were less thoroughly investigated and reviewed. T. 2946 (Garvey). Currently, however, under the new use of force directive, 5002R, all uses of force are reported to the Department's newly created Investigative and Discipline Unit, regardless whether the inmate sustained an injury. Directive 5002R at VIII.

[47] Plaintiffs' Appendix G lists further examples of unusual incident reports in support of plaintiffs' argument regarding defects in the use of force investigations at CIFM.

[48] The May 1988 Use of Force Report concludes that:

In practice, the fact gathering process remains overly dependent on information received from officers directly involved in alleged incidents or from supervisors and managers who have a direct stake in protecting the image of their institutions. The report review process, which relies heavily on the in-house chain of command, does not always resolve factual inconsistencies or discrepancies found in use of force reports. In addition, investigating and drawing conclusions relevant to the use of force is deficient because of the present reliance on inadequately trained Captains and investigators.

p. 41. The Report recommends that special Investigative Assistant Deputy Wardens should assume the investigative duties now carried out by facility Captains. Recommendation # 35.

[49] Despite this shift in responsibility, plaintiffs still argue that their pending motion to add the Inspector General as a defendant should be granted, because the Inspector General will retain responsibility to investigate abuse of inmates by officers when it rises to the level of criminal misconduct. It is concluded, however, that there is insufficient evidence of record concerning the current scope of the Inspector General's role to justify granting this motion at this time. The motion may be revived at a later point if necessary.

[50] A "command discipline" is an "informal, non-adversarial, non-judicial punishment available to a commanding officer to correct minor deficiencies and to maintain discipline within his/her Command." PX 279 (Directive 425R).

[51] See, e.g., Morgan v. District of Columbia, 824 F.2d 1049, 1057 (D.C.Cir.1987); Pressly v. Hutto, 816 F.2d 977, 979 (4th Cir.1987); Zatler v. Wainwright, 802 F.2d 397, 400 (11th Cir.1986); Riley v. Jeffes, 777 F.2d 143, 145-46 (3d Cir.1985); Benson v. Cady, 761 F.2d 335, 339 (7th Cir.1985); Martin v. White, 742 F.2d 469, 473-74 (8th Cir. 1984); Gullatte v. Potts, 654 F.2d 1007, 1012 (5th Cir.1981); Ramos v. Lamm, 639 F.2d 559, 572 (10th Cir.1980), cert. denied, 450 U.S. 1041, 101 S. Ct. 1759, 68 L. Ed. 2d 239 (1981).

[52] The Fourth Circuit has approved a standard which adds to these elements the requirement that the threat of attack result in significant mental pain. Shrader v. White, 761 F.2d 975, 979 (4th Cir.1985). However, no other circuit has followed Shrader's suggestion that the subjective state of mind of inmates at the institution is an element necessary to establish an Eighth Amendment violation.

[53] However, this average figure of "only" 1100 includes only reported incidents of violence, disguises the fact that the rate of violence has been increasing steadily each year, and ignores the fact that many incidents of inmate-inmate violence at CIFM involve more than one inmate.

[54] Plaintiffs' Reply Memorandum of Law at 5. Cf. Alberti v. Klevenhagan, 790 F.2d 1220, 1226 (5th Cir.1986) (commenting on aptness of comparison of number of incidents of violence to the "total annual population of the jail").

[55] Furthermore, plaintiffs have compared the number of yearly documented injuries from inmate-inmate violence at CIFM with the yearly rates of such injuries at several correctional facilities as to which courts have found Eighth Amendment violations. The results indicate that violence occurs at CIFM with comparable frequency. See Plaintiffs' Reply Memorandum of Law at 12 n. 10.

[56] In Rhodes v. Chapman, 452 U.S. 337, 347, 101 S. Ct. 2392, 2399, 69 L. Ed. 2d 59 (1981), the Supreme Court cited the Estelle v. Gamble "deliberate indifference" standard approvingly, going on to state that "[c]onditions [of confinement] must not involve the wanton and unnecessary infliction of pain, nor may they be grossly disproportionate to the severity of the crime warranting imprisonment." See Riley v. Jeffes, 777 F.2d 143, 147 n. 8 (3d Cir.1985). The heightened standard set forth in Whitley v. Albers, 475 U.S. 312, 106 S. Ct. 1078, 89 L. Ed. 2d 251 (1986), to determine officer liability for actions against inmates taken during the course of a prison riot does not apply in an action such as this challenging prison conditions. As the Court of Appeals for the District of Columbia recently noted, "[t]he exigencies and competing obligations facing prison authorities while attempting to regain control of a riotous cellblock" do not exist where the conduct challenged is "the municipality's operation of the Jail generally," because "unlike in the prison riot setting, there can be no legitimate concern that liability will improperly be based on `decisions necessarily made in haste, under pressure, and frequently without the luxury of a second chance.'" Morgan v. District of Columbia, 824 F.2d 1049, 1057-58 (D.C.Cir.1987) (quoting Whitley, 475 U.S. at 320, 106 S.Ct. at 1085).

[57] Although Johnson concerned the rights of pre-trial detainees under the Fourteenth Amendment, other courts have applied the Johnson standard in determining whether the Eighth Amendment rights of convicted prisoners have been violated. See, e.g., King v. Blankenship, 636 F.2d 70, 73 (4th Cir.1980); Furtado v. Bishop, 604 F.2d 80, 95 (1st Cir.1979), cert. denied, 444 U.S. 1035, 100 S. Ct. 710, 62 L. Ed. 2d 672 (1980).

[58] Although, as discussed above, it is impossible to parse out the percentage of the reported uses of force which were legitimate and hence more difficult to compare CIFM to other institutions, the annual number of staff misuse of force findings at CIFM is equivalent to that in the entire 22,000 inmate Los Angeles County jail system. T. 4640 (Painter).

[59] Plaintiffs attach heavy significance to the fact that although fifty-eight officers were initially listed by defendants as witnesses, only fifteen officers actually testified, of which only seven were eyewitnesses to alleged incidents. Plaintiffs ask that the court infer that, had these witnesses testified, their testimony would not have been favorable to defendants, citing Karavos Compania Naviera S.A. v. Atlantica Export Corp., 588 F.2d 1, 9-10 (2d Cir.1978) and Lauratex Textile Co. v. Allton Knitting Mills, Inc., 517 F. Supp. 900, 904 (S.D.N.Y.1981). Defendants vigorously dispute the appropriateness of such an inference, both factually and legally. It is not necessary to resolve this dispute. Both parties were given the opportunity to present evidence concerning incidents of staff-inmate violence. Where defendants chose not to call any eyewitness to challenge an inmate's account of misuse of force and the inmate's account was essentially credible, the weight of the evidence favors the plaintiffs, without the need to draw any inference as a matter of law.

[60] There was also substantial evidence confirming the logical and intuitive proposition that inadequate staffing and overcrowding contribute to disproportionate use of force because officers feel overwhelmed and unable to resolve conflict by any other means.

[61] Cf. Depew v. City of St. Marys, Ga., 787 F.2d 1496, 1499 (11th Cir.1986) (city's continued failure "to take proper remedial action" despite "knowledge of improper police conduct" actionable under § 1983); Villante v. Dept. of Corrections, 786 F.2d 516, 519 (2d Cir.1986) (§ 1983 liability may be based on "an official acquiescence in the continued existence of [unconstitutional prison conditions]"); Fiacco v. City of Rensselaer, N.Y., 783 F.2d 319, 326 (2d Cir.1986) (affirming jury verdict against municipal defendants in police brutality case where evidence was sufficient to show that defendants were "knowingly and deliberately indifferent to the possibility that [their] police officers were wont to use excessive force and that this indifference was demonstrated by the failure of the City defendants to exercise reasonable care in investigating claims of police brutality in order to supervise the officers in the use of force"), cert. denied, ___ U.S. ___, 107 S. Ct. 1384, 94 L. Ed. 2d 698 (1987); Ruiz v. Estelle, 503 F. Supp. 1265, 1302 (S.D.Tx.1980) (liability found where failure to investigate incidents and to train guards had "the practical effect of allowing and even encouraging security officers to indulge in excessive physical violence in the performance of their duties"), aff'd in part and rev'd in part on other grounds, 679 F.2d 1115 (5th Cir.1982).

[62] See Inmates of Occoquan v. Barry, 844 F.2d 828, 841 (D.C.Cir.1988) ("In the prison context, ... district court judges must identify the conditions which, either alone or taken together, violate the Constitution....") (emphasis added); Robles v. Coughlin, 725 F.2d 12, 15 (2d Cir.1983) (complaint alleging inadequate food states a claim under Eighth Amendment); Spain v. Procunier, 600 F.2d 189, 192 (9th Cir.1979) (Eighth Amendment violations found as to use of tear gas and mechanical restraints, and denial of outdoor exercise, although other Eighth Amendment violations denied because "there was no serious deficiency concerning the physical structure and ... medical, nutritional, and sanitary provisions were adequate"); Todaro v. Ward, 565 F.2d 48 (2d Cir.1977) (Eighth Amendment violations found as to certain aspects of medical care).

[63] See also City of Mesquite v. Aladdin's Castle, Inc., 455 U.S. 283, 289, 102 S. Ct. 1070, 1074-75, 71 L. Ed. 2d 152 (1982) (challenge to criminal statute not moot even though challenged language struck by legislature, because city could reenact previous statute if injunction vacated); cf. Huntington Branch, NAACP v. Town of Huntington, 844 F.2d 926, 941 n. 12 (2d Cir.1988) (in case challenging town's refusal to amend zoning plan, court of appeals refuses to take judicial notice of plan to build subsidized housing because, inter alia, "[i]t is entirely speculative and smacks of a mid-litigation effort to demonstrate that the Town is acting in good faith").