Everette SHRADER, Kent Edwin Evans, Merlon Joseph, Dennis Adams, and on behalf of themselves and all others similarly situated, Appellants, and Albert Boisseau, Russell Vinnedge, Plaintiffs, v. Franklin WHITE, Acting Director; Robert M. Landin, Acting Director; Terry C. Richtmeyer, Regional Administrator; Elwood Booker, Superintendent; Rufus Fleming, Assistant Superintendent; Edward Wright, Institutional Security Chief, Appellees

CHAPMAN, Circuit Judge:

The plaintiffs-appellants, inmates at Virginia State Penitentiary (VSP), brought this 42 U.S.C. § 1983 class action for declaratory and injunctive relief alleging that various conditions of their confinement violated the eighth amendment’s proscription against cruel and unusual punishment. The parties consented to trial before a United States magistrate, who concluded that the conditions at VSP do not violate the eighth amendment and dismissed the inmates’ complaint. On appeal the inmates assert that the magistrate applied an incorrect legal standard with respect to obtaining injunctive and/or declaratory relief from the threat of violence and sexual assault from other inmates at VSP. The inmates also assert that the magistrate made clearly erroneous findings of fact regarding inmate safety, the physical plant, fire hazards, and food service at VSP. Finding that the magistrate applied the correct legal standard and that his findings of fact were not clearly erroneous, with one exception, we affirm.

I

Plaintiffs initially assert that the magistrate applied an erroneous legal standard for injunctive and/or declaratory relief from the threat of violence and sexual assault from other inmates at VSP. Plaintiffs argue that the magistrate’s standard ignores this court’s holdings in Woodhous v. Virginia, 487 F.2d 889, 890 (4th Cir. 1973), and Withers v. Levine, 615 F.2d 158 (4th Cir.1980), cert. denied, 449 U.S. 849, 101 S.Ct. 136, 66 L.Ed.2d 59 (1980). The defendants assert that the standard used by the magistrate is a proper application of Woodhous and Withers in light of Rhodes v. Chapman, 452 U.S. 337, 101 S.Ct. 2392, 69 L.Ed.2d 59 (1981). We agree.

Specifically, the inmates assert that the magistrate erred in requiring that inmates suffer “significant mental pain” and that the danger be unnecessarily and wantonly inflicted and “totally without penological justification” to establish prison violence of constitutional proportions. As will be shown, however, the magistrate correctly applied the eighth amendment standard established in this Circuit and the Supreme Court. In Woodhous v. Virginia, 487 F.2d at 890, we set forth a standard of liability regarding unconstitutional levels of prison violence:

While occasional, isolated attacks by one prisoner on another may not constitute cruel and unusual punishment ... confinement in a prison where violence and terror reign is actionable. A prisoner has a right, secured by the eighth and fourteenth amendments, to be reasonably protected from constant threat of violence and sexual assault by his fellow inmates, and he need not wait until he is actually assaulted to obtain relief.

******

In determining whether to grant relief, the court should ascertain: (1) whether there is a pervasive risk of harm to inmates from other prisoners, and, if so, (2) whether the officials are exercising reasonable care to prevent prisoners from intentionally harming others or from creating an unreasonable risk of harm, (citations omitted).

Later, in Withers v. Levine, 615 F.2d at 161, we elaborated on our holding in Wood-hous as follows:

A pervasive risk of harm may not ordinarily be shown by pointing to a single incident or isolated incidents, but it may be established by much less than proof of a reign of violence and terror in the particular institution. The defendants seized upon that explanatory phrase from Woodhous to contend that something approaching anarchy must be proven before a cause of action under Wood-hous may be made out, but conditions need not deteriorate to that extent before the constitutional right to protection arises. It is enough that violence and sexual assaults occur ... with sufficient frequency that the younger prisoners, particularly those slightly built, are put in reasonable fear for their safety and to reasonably apprise prison officials of the existence of the problem and the need for protective measures. ...

It is not necessary to show that all prisoners suffer a pervasive risk of harm. It is enough that an identifiable group of prisoners do, if the complainant is a member of that group.

In considering “for the first time the limitation that the Eighth Amendment ... imposes upon the conditions in which a State may confine those convicted of crimes,” Rhodes, 452 U.S. at 344-45, 101 S.Ct. at 2397-98, the United States Supreme Court acknowledged that the eighth amendment’s protections reach “beyond the barbarous physical punishments at issue in the Court’s earliest cases.” Id. at 345, 101 S.Ct. at 2398. The Court announced:

Today the Eighth Amendment prohibits punishments which, although not physically barbarous, “involve the unnecessary and wanton infliction of pain,” Gregg v. Georgia, [428 U.S. 153, 173, 96 S.Ct. 2909, 2925, 49 L.Ed.2d 859 (1976)], or are grossly disproportionate to the severity of the crime, Coker v. Georgia, 433 U.S. 584, 594 [97 S.Ct. 2861, 2867, 53 L.Ed.2d 982] (1977) (plurality opinion); Weems v. United States, 217 U.S. 349 [30 S.Ct. 544, 54 L.Ed. 793] (1910). Among “unnecessary and wanton” inflictions of pain are those that are “totally without penological justification.” Gregg v. Georgia, supra, at 183 [96 S.Ct. at 2929]; Estelle v. Gamble, 429 U.S. 97, 103 [97 S.Ct. 285, 50 L.Ed.2d 251] (1976).

452 U.S. at 346, 101 S.Ct. at 2399 (footnote omitted).

The Court went on to discuss how courts should decide whether conditions violate the eighth amendment:

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.” Trop v. Dulles, 356 U.S. 86, 101 [78 S.Ct. 590, 598, 2 L.Ed.2d 630] (1958) (plurality opinion). The Court has held, however, that “Eighth Amendment judgments should neither be nor appear to be merely the subjective views” of judges. Rummel v. Estelle, 445 U.S. 263, 275 [100 S.Ct. 1133, 1139, 63 L.Ed.2d 382] (1980).

452 U.S. at 346, 101 S.Ct. at 2399.

The plaintiffs challenge the following portion of the magistrate’s discussion of the law regarding “mental pain”:

One key to understanding when the risk of violence reaches constitutional dimensions is its effect on the inmate population ____ In this context, it is not necessary that an inmate establish that he has been the subject of an actual attack, but he must establish that he lives in reasonable fear of assaults from other inmates ... and that the fear results in significant mental pain ____ To establish fear of constitutional dimensions, an inmate must show more than simple anxiety. He must demonstrate anxiety on a level such as would interfere to some degree with his everyday functions____ An inmate does not need to establish that he is totally incapacitated by any means. But, before pain of a constitutional magnitude can be said to exist, there must be evidence of serious mental and emotional deterioration attributable to the fear of constant danger from assaults.

Certainly it is not improper to consider the effect of conditions on the prison population in determining if the conditions are of constitutional magnitude. Rhodes v. Chapman, 452 U.S. at 364, 101 S.Ct. at 2408 (Brennen, J., concurring). We did as much in Sweet v. South Carolina Dept. of Corrections, 529 F.2d 854, 865-866 (4th Cir.1975), when we remanded the case for a determination of whether an indefinite limitation on exercise is harmful to a prisoner’s health, and in Johnson v. Levine, 588 F.2d 1378, 1380 (4th Cir.1978), where we affirmed the district judge’s findings of unconstitutional overcrowding, noting that the overcrowding had probably contributed to, inter alia, “psychological injury to some prisoners.” In addition, the magistrate correctly noted that to obtain relief an inmate need not be the victim of an actual attack, but rather must suffer from the reasonable fear of inmate assaults. Withers, supra; Woodhous, supra. Since under Rhodes conditions at the prison which, “although not physically barbarous, involve the unnecessary and wanton infliction of pain” are prohibited under the eighth amendment, Rhodes, 452 U.S. at 346, 101 S.Ct. at 2399 (quoting Gregg v. Georgia, 428 U.S. 153, 173, 96 S.Ct. 2909, 2925, 49 L.Ed.2d 859 (1976)), the magistrate’s requirement that fear of attack re-suit in significant mental pain to be of constitutional dimensions is proper.

The plaintiffs also challenge the magistrate’s statement that the fear amounting to mental pain must be inflicted recklessly or wantonly and be “totally without peno-logical justification.” That portion of the magistrate’s order is as follows:

Nevertheless, assuming that an inmate establishes that his fear of assaults has produced significant mental distress and psychological impairment to produce “pain,” it does not automatically follow that he has been subjected to cruel and unusual punishment. He must also show that the pain was inflicted recklessly or wantonly____ Moreover, the pain must, itself, be “totally without penological justification.” (citations omitted).

Both of these standards are set forth in Rhodes v. Chapman, 452 U.S. at 346, 101 S.Ct. at 2399. The inmates argue that requiring them to show reckless or wanton infliction of pain ignores our decision in Withers, which determined that negligence by prison officials is a violation of the constitutional right and actionable under § 1983. That determination was not without some qualification, however. In Withers we recognized that:

Surely the Constitution of the United States does not lend its protection to every victim of a common law tort by state officials and employees ... [b]ut negligence by a state official under some circumstances may itself violate a constitutionally protected right; when it does, it is actionable under § 1983____

When there is present in a prison or in an identifiable portion of it, a pervasive risk of harm to all prisoners, or to an identifiable group of them, the constitutional prohibition against cruel and unusual punishment requires that prison officials exercise reasonable care to pro vide reasonable protection from such unreasonable risk of harm.

Withers, 615 F.2d at 162. The plaintiffs acknowledge that the Supreme Court stated in Rhodes v. Chapman that the infliction of pain must be unnecessary and wanton to constitute a violation of the eighth amendment, but argue that we should apply our Withers reasonable care standard rather than the Supreme Court’s unnecessary and wanton standard because Rhodes v. Chapman involved only the issue of double celling and the principles it set forth should only be applied within that narrow context. But even a cursory reading of Rhodes reveals its discussion of “unnecessary and wanton infliction of pain” is not so limited: “These principles apply when the conditions of confinement compose the punishment at issue.” 452 U.S. at 347, 101 S.Ct. at 2399 (emphasis added). We thus decline to ignore the mandate of the United States Supreme Court that the eighth amendment prohibits punishments which “involve the unnecessary and wanton infliction of pain.” The magistrate properly applied the principles of Withers and Woodhous in the light of Rhodes v. Chapman in adopting the legal standard he applied to the evidence in this case.

II

On the issues of inmate safety relating to weapons, drugs, cell locking system, cell house security, manpower, idleness and protective custody, the appellants’ brief is so filled with exaggerated statements and claims that we thought they were hyperbole. At oral argument it became obvious that we were expected to accept these statements as true. An examination of the record fails to support such exaggeration of the claims.

Our role in reviewing the findings made by the lower court is to determine whether they are clearly erroneous, for “[o]nly when a finding of fact is ‘clearly erroneous’ are we to set it aside.” Phillips v. Crown Central Petroleum Corp., 556 F.2d 702, 703 (4th Cir.1977). “A finding is ‘clearly erroneous’ when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.” United States v. Gypsum, 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948). In reaching his findings the magistrate noted that “there is conflicting evidence involved in every factual discussion of these issues. As such, the resolution of these issues must be determined by weighing the credibility of the evidence presented by the concerned parties.” As an appellate court, we must “accord wide latitude to findings of the trial court with respect to credibility of witnesses, and to facts and to inferences therefrom.” Phillips, 556 F.2d at 703. We reject at the outset the inmates’ contention that the considerable documentary evidence in this case was “largely overlooked by the magistrate.” The record reveals that the magistrate read all the exhibits and reviewed them before rendering his decision. (App. 33).

The men confined at YSP have been convicted of crimes of violence. They are confined against their will. Under such circumstances acts of violence by inmates against inmates are inevitable. No amount of money and no increase in the number of prison officers is going to completely eradicate inmate violence from VSP or any other such institution. This is a maximum security prison and each inmate has been interviewed, tested and classified before being confined to VSP. Often an effort to achieve the “evolving standards of decency that mark the progress of a maturing society,” Trop v. Dulles, 356 U.S. 86, 101, 78 S.Ct. 590, 598, 2 L.Ed.2d 630 (1958), runs headlong into the hardened criminals who are incarcerated in our maximum security prisons. Methods used by prison administrators to protect inmates from one another are often attacked as oppressive or as constituting cruel and unusual punishment by the same inmates who later complain about a lack of protection. To follow the plaintiff’s argument we would have to find that imprisoned felons are constitutionally entitled to more protection from crimes of violence than are law abiding citizens.

The plaintiffs complain of unprovoked assaults, but the rate of inmate assaults at VSP compares favorably to Lucasville Prison which was the subject in Rhodes. In any event, the important point is that the magistrate found, and his finding is amply supported by the record, that “the evidence of any real fear of assaults at Virginia State Penitentiary is virtually nonexistent.” The plaintiff Joseph had been incarcerated at VSP for eight years and had never been assaulted and testified that he had no fear of assault. Plaintiff Evans has been a prisoner for six years, has never been physically struck, and was not afraid of physical danger. Plaintiff Vinnedge had been confined for eight years and was serving his second assignment to VSP at his own request. Plaintiff Shrader had been confined at VSP for eleven years and been robbed once, but never physically injured by another inmate. The inmates who testified they had been assaulted included persons who had instigated fights or were aggressors. One was found to be in possession of marijuana and another was operating as a “storeman” which is comparable to a black market operation in which goods are sold to other inmates for a profit. The officials had tried to stop this activity and after learning of a threat against the “storeman” he was put into protective custody. The inmate-on-inmate assault rate at VSP is lower than that in some modern prisons and the rate compares favorably with other institutions.

Only one homosexual attack has been reported since 1979. The magistrate found, and his finding is amply supported by the record that VSP does not have a large number of young inmates. Appellants cite as youths one inmate 22 years of age and another of 23 at VSP. However, one of these inmates is under a sentence of capital murder, and the other is serving 30 years and has incurred more than 20 disciplinary charges at VSP.

There is a problem of illegal drug use at VSP, but it involves only a small percent of the inmates. Most of the incidents of illegal drug use involve Talwin, but this drug has since been removed from the infirmary. Risk of exposure to illegal drugs inside VSP seems no greater than the risk of exposure on the outside.

The magistrate’s finding that the conditions of protective custody and administrative segregation meet constitutional standards is amply supported by the record. Although these prisoners are confined to their cells, they are permitted radios and reading materials, have the same meals as the general prison population, and are allowed limited commissary privileges. They are provided regular times to bathe and to exercise. The cells used for protective custody and administrative segregation are adequately heated, lighted and ventilated. When an inmate is in protective custody at his own request he is not subjected to cruel and unusual punishment simply because he may be deprived of certain privileges available to the general prison population. Breeden v. Jackson, 457 F.2d 578 (4th Cir.1972).

Appellants complain of a lack of security manpower to properly staff VSP. The record reveals that there are 307 security personnel for approximately 900 inmates, a ratio of 3:1, which is much better than the national average of 5:1. The record supports the finding that the security staffing was “completely adequate.”

There is some idleness among VSP inmates, but this is of their own choosing. Inmates have the opportunity to work at prison jobs or to enroll in various programs. Inmates who are motivated can enroll in many activities and programs. Appellants’, witness Conquest complained that his job took very little of his time, but on cross examination he admitted that he spent 15 hours per week in a radio/TV repair class, lifted weights each day, participated in the VSP drug program, participated weekly in a VSP theatre group, and was considering a computer course and college correspondence courses. There are 178 educational and vocational positions for the inmates and additional positions can be created when there is a sufficient inmate interest. There is no “hard labor” at the prison and inmates are normally confined to their cells only when sleeping, and they are free to participate in the various vocational educational and recreational activities.

The complaints about the locking system of the cells of A building do not create a constitutional problem. This system is old and needs frequent repair, but repairs are promptly made when there is a malfunction. This locking system locks or unlocks all doors at the same time. Any locking system for prison cells will produce complaints from inmates who might prefer something different.

The presence of weapons in the institution concerns us, particularly in view of the magistrate’s reasoning. He stated in footnote 32:

Plaintiffs have stressed the prevalence of weapons in the Virginia State Penitentiary by claiming that the state has failed to protect them from a pervasive risk of harm, relying on the language of Withers, 615 F.2d 158. At first glance, the argument is attractive, for weapons do exist. There is evidence that the material from which they were made, i.e., scrap metal, was not sufficiently safeguarded and there is evidence that the method employed to restrict the making and conveyance of such weapons was inadequate. The facts may well establish that a danger exists which is specific (the weapons), obvious (from the number of stabbings and shakedown reports), and “unnecessary.” The fallacy in the argument is the use of the term “unnecessary.” The danger does not come from the weapons themselves. It comes from the use of those weapons, and as has already been discussed, that risk, in terms of an Eighth Amendment violation, is not unnecessary. The weapons may increase the risk of violence, but their existence is only one facet to be considered in the overall question of whether or not prison officials have acted in a willful and wanton manner in inflicting unnecessary pain. The fact that inmates have been able to make and possess some knives out of scrap metal may be impressive. It may even seem shocking. But, the stark reality of prison life is that virtually anything can be employed as a weapon, e.g., a pencil can become a dagger; an electric cord a garrote, a lock inside a sock a bludgeon; human excretion a poison. Pool cues, brooms, and chairs have all been used as weapons at one time or another. The pervasive risk of harm from which inmates are entitled to be protected is not the knife, but the act of assault. The courts have been unwilling to establish constitutional liability solely on the basis of the existence of weapons within an institution. See Bogard v. Cook, 586 F.2d 399, 418 (5th Cir.), cert. denied, 444 U.S. 883 [100 S.Ct. 173, 62 L.Ed.2d 113] (1979); cf. Puckett v. Cox, 456 F.2d 233, 235 (6th Cir.1972) (negligence in permitting insane prisoner to roam within prison, and in allowing him access to dangerous instruments does not state equal protection claim).

While a knife is not dangerous in and of itself, and only becomes dangerous when used to inflict injury, we are troubled by the statement that scrap metal is not sufficiently safeguarded and that the methods employed to restrict the making and conveyance of weapons are inadequate. This would indicate a pervasive risk of harm from use of the knives, and a possible inadequate official response would appear to make out a claim if prison officials ignore the problem, after they become aware of it, and fail to take action to prevent the manufacture and possession of weapons which can be used to harm other inmates. Although the evidence does not establish that there have been many injuries from these weapons, the condition should not be allowed to exist if it can be prevented by better inventory control and supervision of the machine shop or other places where scrap metals are produced and can fall into the hands of inmates.

We cannot determine from the magistrate’s findings whether the manufacture of weapons from scrap metal is being con ducted on such a scale as to represent a pervasive risk of harm or what the official response has been.

Ill

The inmates also assert that the magistrate erred in finding that the physical plant and food service are constitutionally adequate and that there is no unreasonable risk of harm from fire or smoke.

Physical Plant

Regarding the condition of the physical plant at VSP, the magistrate noted:

Inmates confined in the Virginia State Penitentiary are furnished a clean cell for single occupancy, with adequate heat and ventilation. Each inmate has his own toilet and sink. Although an inmate may only have immediate access to cold water inside his cell, he can readily obtain hot water for cleaning and personal hygiene. Adequate showers are available on each tier and freedom of movement in each section is relatively unham-pered____

The magistrate also stated that “the buildings cannot be described as cheerful.” The court recognized that “the facilities at the Penitentiary are not new, but age alone is far less important than the state of repair of a building in considering whether conditions of confinement offend the Eighth Amendment.” The magistrate concluded that “the physical conditions surrounding the plaintiffs’ confinement in the Virginia State Penitentiary, either separately or in combination, do not amount to the infliction of cruel and unusual punishment.”

The plaintiffs argue that “the magistrate failed to mention, and apparently failed to consider, the extensive and unrefuted evidence from defendants’ experts and defendants’ own documents that the Penitentiary was in such a state of disrepair that it should be closed.” However, the defense expert referred to by the plaintiffs, Terrell Don Hutto, a criminal justice consultant, testified that operations should be “phase[d] out” at the Penitentiary because “it [is] old and economically inefficient for the Commonwealth.” Such testimony is a far cry from stating the Penitentiary should be closed because it is in “such a state of disrepair.” The plaintiffs also assert that “Mr. Clarke, Penitentiary Building Superintendent confirmed the major and persisting maintenance and structural problems.” The record reflects that Mr. Clarke testified not only as to the ongoing maintenance problems, but also as to the ongoing repairs made at VSP and the current plans for additional renovations at VSP “to the cell blocks and to the kitchen refrigeration.” Mr. Clarke described his maintenance program as “not a complete functioning program. We do preventive maintenance.”

The plaintiffs also challenge the magistrate’s findings with respect to leaks in the roof in A building and the resultant water damage to the building. The magistrate observed minor wetness under the windows in the building, which appeared to come from the windows and not the roof, and found that it has “not created any safety or health hazards.” One of the defendants’ witnesses testified that on occasion he “ha[d] seen puddles” in A-cell house. He agreed that in certain areas this was because of some ceiling deterioration due to water. Defense witness Bill Davis, who inspected the roof and ceiling of A building, found no structural damage and only one area where the metal ceiling had rusted through. This area had been repaired at the time of his inspection. Moreover, Davis was in A building when it was raining and saw no rain leaking into the building. Also, immediately after one inmate testified that “it’s raining inside A building right now” the magistrate recessed, toured A building, and found only minor wetness under some windows. Our review of the record reveals the magistrate’s findings with respect to water seepage and damage not reaching constitutional proportions are not clearly erroneous.

Regarding inmates’ complaints of pigeons and other birds nesting in A building, the magistrate found that the pigeon problem has been eliminated because the cupolas in the roof have been screened. The magistrate also found no evidence that any of the birds created a health problem for the inmates. These findings are fully supported in the record. Defense witness Mr. Clarke testified that although attempts in the past to keep the pigeons out had not been successful, screening the cupola in the attic has effectively prevented the entry of pigeons. Inmate Conquest also testified that the birds are no longer a problem “because they patched the holes in the ceiling.”

The plaintiffs assert that their health expert, Mr. Gordon, gave undisputed testimony about the serious health risks posed by pigeon feces. Mr. Gordon did testify that “pigeon feces and pigeons inside of the building increase the opportunity for the occurrence and transmission of certain mycotic diseases.” Defendants’ health expert, on the other hand, found nothing that disturbed him from a sanitation point of view in the cellblocks. No evidence was introduced to show any inmate had suffered ill effects from the presence of pigeons in the past.

With respect to the showers at VSP, the magistrate found that they leave much to be desired____ The show-

er heads drip and rust stains are plentiful. There is evidence that mildew or mold may also be present. One inmate indicated that there might be a momentary loss of cold water if a nearby toilet were to be flushed. There is no evidence that any inmate was ever injured from a malfunction of the showers, nor is there any evidence that the conditions of the stalls created a health hazard. One might have expected at least one complaint of athlete’s foot, but there were none.

The magistrate concluded the showers at VSP were adequate, which is fully supported by the record. Although inmates Collins and Crute testified that the hot and cold shower controls frequently do not work well, neither that condition nor the momentary loss of cold water in the showers when a toilet is flushed are of constitutional significance. “[A] sentence in federal prison is not a guarantee that one will be safe from life’s occasional inconveniences.” Holloway v. Gunnell, 685 F.2d 150, 156 (5th Cir.1982). The defendants concede that mold in some showers is a problem; inmates are paid to clean the cell houses, including the showers, and some are cleaned more thoroughly than others. In light of the evidence of daily attempts to clean the showers and no evidence of disease resulting from mold in the showers, the conditions of the showers at VSP do not, alone or in combination with other conditions, constitute cruel and unusual punishment.

Finally, the inmates complain that “the magistrate ignored both plaintiff and defendant expert testimony as well as defendants’ own documents which confirm that cell lighting is substandard and can cause health problems.” To the contrary, the magistrate specifically noted that “technical measurements of the interior cell lighting revealed the illumination to be approximately one-fourth to one-half that recommended by the American Correctional As-sociation____” Yet the magistrate found the cell lighting to be “relatively good” based on his unannounced toiir of VSP at 10:00 p.m. The magistrate stated, “Many inmates were reading or writing letters; some were typing material for further consideration by this court; and, others had extinguished their lights to improve their television viewing.” In addition, technical measurements of the lighting at the Bench, at counsel tables, and at the clerk’s desk in the courtroom where this case was tried revealed lighting comparable to that in the cells at VSP. Considering the record, the illumination in the cells at VSP “does not offend evolving standards of decency.” Rhodes v. Chapman, 452 U.S. at 346, 101 S.Ct. at 2399.

Fire Safety

The magistrate found that the plaintiffs are not exposed to an unreasonable risk of harm by fire or smoke. The record fully supports this finding. The record reveals that the cell buildings at VSP are constructed primarily of steel and concrete which does not burn easily. The only wood in A building are 2x4 boards framed between steel beams used to support a metal ceiling and a four-foot wooden catwalk between the steel beams supporting the roof. Testimony of the defendants’ expert established that due to the height of the ceiling, a cell fire would not cause combustion of any of the wood. While flammable items are kept by the inmates in their cells, the number of these items has been reduced by regulation at the Penitentiary. In addition, all flammable mattresses in the cells at VSP have been replaced with fire-resistant boric acid treated mattresses. The plaintiffs challenge the court’s specific finding that “the evidence establishes that a fire in any given cell would be unable to spread to an adjacent cell____” The plaintiffs point out that two of their experts, Mr. Silver and Mr. Whitfield, testified that a fire of combustibles within a cell could create combustion in an adjacent cell. The record reveals that there is conflicting testimony on this point. The defendants’ expert testified that even under the worst possible conditions, including the addition of kerosene, it was unlikely that a fire in one cell would be hot enough to spread to an adjoining cell. The magistrate took note of the conflicting testimony and “place[d] its confidence” in the defendants’ expert who, unlike the plaintiffs’ expert, stated the specific facts on which he based his expert opinion. The magistrate’s finding that the risk of fire in the cellhous-es is negligible is not clearly erroneous.

The plaintiffs also assert that the magistrate erred in finding that “the volume of space in each of the cellhouses lends credence to the ... testimony indicating that the smoke and fumes would rise relatively harmlessly to the ceiling and then disburse.” The plaintiffs point to the testimony of their expert, Mr. Silver, that smoke from a fire forms a conical pattern and rises upward, which could cause the inmates on the upper tiers of the building to be “buried in an opaquing environment.” On the other hand, the Fire Safety Officer at VSP testified that smoke could rise freely into an open area above the ceiling and below the roof and then be ejected by fans located in A and B buildings. The magistrate found that “the volume of space in each of the cellhouses lends credence to the later testimony indicating that the smoke and fumes would rise relatively harmlessly to the ceiling and then disburse.” Since he was able to judge the credibility of the witnesses and had inspected the facility, we decline to disturb this finding on appeal.

Finally, the inmates challenge the magistrate’s assertion that “it is difficult to believe that anyone would deliberately expose himself to a great risk of harm from excessive smoke and fumes by committing arson.” They argue that the court failed to credit the testimony of their expert, Mr. Silver, who testified that 96% of fires started in correctional facilities are arson, thus requiring greater precaution than in a civic building. However, the lower court noted that he did not ignore that testimony, nor did he ignore the fact that arson has occurred at VSP. Rather, he found it “difficult to believe.” “Accordpng] wide latitude to the findings of the trial court with respect to credibility,” Phillips, 556 F.2d at 703, we do not find it clearly erroneous that the trial court had difficulty believing that inmates would deliberately injure themselves by arson. Many prison arsonists are not trying deliberately to harm themselves, but are trying to call attention to themselves or to some grievance they may have with the prison administration.

The record reveals that deficiencies in fire safety do exist at VSP, and that some of the problems, such as the single means of egress from each cellhouse, violate Virginia Fire Safety Regulations. However, as noted by the magistrate, such “safety codes do not establish the constitutional minima,” Ruiz v. Estelle, 679 F.2d 1115, 1153 (5th Cir.1982) (quoting Bell v. Wolfish, 441 U.S. 520 at 543 n. 27, 99 S.Ct. 1861 at 1876 n. 27, 60 L.Ed.2d 447 [1979]), which are “whether the actual conditions of confinement ... are cruel and unusual.” Nelson v. Collins, 659 F.2d 420, 426 (4th Cir. 1981). The deficiencies in fire safety at VSP, whether considered alone or in combination with other conditions, do not constitute cruel and unusual punishment.

Food Service

It is well-established that inmates must be provided nutritionally adequate food, “prepared and served under conditions which do not present an immediate danger to the health and well being of the inmates who consume it.” Ramos v. Lamm, 639 F.2d 559, 571 (10th Cir.1980), cert. denied, 450 U.S. 1041, 101 S.Ct. 1759, 68 L.Ed.2d 239 (1981); Bolding v. Holshouser, 575 F.2d 461 (4th Cir.1978), cert. denied, 439 U.S. 837, 99 S.Ct. 121, 58 L.Ed.2d 133 (1978). In the instant case the magistrate ruled that plaintiffs did not prove their claims of unsanitary food conditions or nutritional inadequacies at VSP. The parties stipulated that the cyclical menu at VSP, if followed with properly prepared and stored food, provides adequate nutritional diets. The parties disagree about the way the food is stored, prepared, and served. The magistrate observed:

The evidence is in conflict in every area. If one were to believe the plaintiffs’ inmate witnesses, rarely a meal is served that does not consist of: (1) raw food; (2) rotten food; or (3) contaminated food ... The defendants presented evidence of a Four-Star restaurant with excellent management procedures. In the final analysis, it all boils down to one thing— credibility.

The plaintiffs’ nutritional expert Wilson observed various unsanitary conditions on her visit to the VSP mess hall area, but the magistrate noted that those conditions were atypical of the true conditions at VSP as described by persons who regularly visit the penitentiary. Testimony of the Food Service Manager and an inmate indicates the inmates who work in the kitchen and who knew of Ms. Wilson’s visit 24 hours before she arrived intentionally made the kitchen look dirty and in a hectic condition that day. Ms. Wilson herself admitted that some unsanitary conditions, such as clogged sinks, had disappeared when she returned to VSP.

The magistrate acknowledged that there were violations reported in the state sanitation reports, but found insufficient evidence that the problems persisted continuously without corrective action. Two of the defendants’ experts testified that the ongoing extermination program at VSP controlled mice to the extent that they did not present a real public health problem.

The record also reveals no evidence of outbreaks of food poisoning, diarrhea, or other diseases which are indicative of unhealthy conditions in the preparation or handling of food at VSP. Several inmates testified they felt discomfort after eating some meals, but only two said they had ever sought medical attention for such discomfort. Moreover, the chief physician for the Department of Corrections reviewed the records of all inmates who testified as having food-related complaints and found that no two complaints occurred on the same day, indicating “there wasn’t any common factor making these individuals sick.” Furthermore, the magistrate noted there is no evidence of either malnutrition or “sub-clinical manifestations” of malnutrition, such as increased dental problems, loss of appetite, reduced capacity to work, ward off disease, cope with stress, etc. Plaintiffs’ nutritional expert, Ms. Wilson, testified she would have expected the conditions at VSP to result in such “sub-clinical manifestations” of malnutrition. The magistrate’s findings regarding the conditions of the food service area at VSP were not clearly erroneous.

In Rhodes v. Chapman, 452 U.S. 337, 101 S.Ct. 2392, 69 L.Ed.2d 59 (1981), the Supreme Court provides guidance to judges in eighth amendment claims arising from prison conditions. "In 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 detention facility.’ ” Id. at 351, 101 S.Ct. at 2401 (quoting Bell v. Wolfish, 441 U.S. at 539, 99 S.Ct. at 1874).

Rhodes also reminds us, “The Constitution does not mandate comfortable prisons, and prisons of SOCF’s [Southern Ohio Correctional Facility’s] type, which houses persons convicted of serious crimes, cannot be free of discomfort.” Rhodes, 452 U.S. at 349, 101 S.Ct. at 2400.

Applying the Rhodes standard, we affirm all of the district court’s findings and conclusions except as to the construction of weapons from scrap metal and remand this issue to the district court for additional proceedings to determine the extent of this danger, the defendants’ response thereto, and whether this danger and response or lack of response rise to the level of an eighth amendment violation.

AFFIRMED IN PART.

REMANDED IN PART.

. The district court certified the case as a class action on behalf of all inmates incarcerated, or to be incarcerated, at the facility.

. It should be noted that the lower court found the evidence of any real fear of assaults at VSP to be virtually nonexistent.” For example, the court found that: On each occasion that the court visited the Penitentiary, there was a striking lack of tension. On the unannounced tour, it appeared that the inmates were relaxed and totally unconcerned that they might be subjected to any violence.”

. Actually, only one of the two experts opined with any certainty that a fire in one cell could produce enough heat to cause the adjoining cells to ignite. The plaintiffs’ other fire expert, J.M.M. Whitfield, testified, "I guess there is always a possibility that flash point could be reached from one cell to another, but I am not sure how likely that would be.”

SPROUSE, Circuit Judge,

concurring in part and dissenting in part:

I respectfully dissent in part.

Judge Chapman as usual has crafted a well-written opinion. I agree with and concur in what he has written in Parts I and III of his opinion regarding the legal standard to be applied in this case and the factual findings concerning the physical plant, the food service, and fire safety, but I feel that the magistrate was clearly erroneous in some of the factual findings discussed in Part II of the majority opinion. I would reverse as clearly erroneous the magistrate’s factual findings concerning the risk of violence as a result of the prevalence of weapons and drugs in the prison, the level of the inmates’ fear of that violence, and the adequacy of security precautions in suppressing violence.

Prior to setting out his findings of fact, the magistrate noted that the voluminous evidence was in sharp conflict and that resolution of the issues in the case “must be determined by weighing the credibility of the evidence.” In his memorandum opinion, however, he recorded only two specific credibility findings related to inmate safety. Credibility determinations of a fact finder of course will be “aeeord[ed] wide latitude” by an appellate court. Phillips v. Crown Central Petroleum Corp., 556 F.2d 702, 703 (4th Cir.1977); 9 Wright & Miller, Federal Practice and Procedure § 2586, at 736-37 (1971). A finding may not be insulated from review, however, by vague statements that credibility issues are presented by some undesignated testimony. Such a rule would totally defeat appellate review of factfinding upon the finder’s mere recitation that there are credibility conflicts — here, for example, scattered over eleven volumes of trial transcript and hundreds of exhibits. In my view we should be guided in this case by the general “clearly erroneous” standard of review, tempered by the “credibility” rule only in those two instances when the magistrate made specific credibility determinations. Under the clearly erroneous standard, in order to reverse rulings based on findings of fact, we must conclude that although the record contains evidence to support the findings, our review of the entire record leaves us with the definite and firm conviction that a mistake has been committed. United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948). Such conviction may follow when the findings were induced by an erroneous view of the controlling legal standard, are not supported by substantial evidence, do not account for substantial evidence to the contrary, or are against the clear weight of the evidence considered as a whole. Miller v. Mercy Hospital, Inc., 720 F.2d 356, 361 (4th Cir.1983).

Under that standard I cannot say that the magistrate committed clear error in his findings relating to the physical plant, the food service, and fire safety, so no purpose would be served by discussing that evidence. Calling for reversal in this case due to erroneous factual findings relating to the risk of harm to the inmates from violence, however, necessarily entails a particularized recitation of facts, considering the evidence on both sides of the issues. The majority’s synthesis of Rhodes v. Chapman, 452 U.S. 337, 101 S.Ct. 2392, 69 L.Ed.2d 59 (1981), Withers v. Levine, 615 F.2d 158 (4th Cir.), cert. denied, 449 U.S. 849, 101 S.Ct. 136, 66 L.Ed.2d 59 (1980), and Woodhous v. Virginia, 487 F.2d 889 (4th Cir.1973), correctly demonstrates that to show an eighth amendment violation, the inmates must prove 1) a pervasive risk of harm from violence, 2) significant mental pain caused by the inmates’ fear of harm from violence, and 3) “unnecessary and wanton infliction” of that pain by the deliberate or reckless failure of prison officials to take adequate precautions to reduce the risk of violence. The magistrate found “ample evidence of a number of serious injuries” from inmate-on-inmate assaults, but he found further that significant mental pain was absent because the “evidence of any real fear of assaults ... is virtually nonexistent” and that the state “has recognized the dangers” and has taken reasonable and adequate steps “to reduce the risk of violence.” In my view, however, the trial evidence clearly requires a finding in favor of the inmates on each of these points.

I. Pervasive risk of harm

In order to establish a violation of the eighth amendment, the inmates’ trial burden was to prove that there exists at VSP a pervasive risk of harm from violence amounting to more than isolated incidents, but not necessarily rising to the level of a reign of violence. 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). The incidence of violence must be sufficient to put inmates “in reasonable fear for their safety and to reasonably apprise prison officials of the existence of the problem and the need for protective measures.” Id. In my view the prisoners met that burden by showing that a pervasive risk of harm exists because of the frequency and brutality of inmate-on-inmate assaults which result directly from the easy access to contraband weapons and drugs inside VSP. The magistrate reviewed the official statistics on assaults and conceded that in VSP the inmates face a “high risk” of violence, but his findings of insufficient mental pain and adequate security precautions apparently obviated any need for more detailed findings. The evidence, however, that the inmates at VSP suffer a pervasive risk of harm from inmate-on-inmate violence is compelling.

Statistics recited by the magistrate indicate that VSP is a highly dangerous place for inmates: They show seven inmates murdered over the five-year period preceding trial, as well as fifty-four stabbings and twenty-four other serious inmate-on-inmate assaults. Other evidence at trial indicates, moreover, that these statistics, apparently tabulated directly from official Serious Incident Reports, significantly understate the true frequency of violent assaults at VSP. Former assistant warden Reynolds, testifying on direct examination by defendants, reported a total of ninety-six inmate-on-inmate assaults during the three-year period 1980-82. The inmates’ corrections expert witness Dr. Fogel testified that the prison’s Serious Incident Reports understate the frequency of assault because incidents in which more than one inmate is injured are reported as a single incident, and some incidents are reported as “other” or of unexplained nature when they are clearly inmate-on-inmate assaults. As an example, VSP corrections sergeant Hotchkiss, testifying for the inmates, recounted an instance when a brawl between inmates over which of them would have another inmate as a “punk” (apparently a homosexual partner) was reported as a slip and fall in the shower. Dr. Fogel analyzed the Serious Incident Reports and recomputed the figures to correct this misreporting problem, and he calculated a total of ninety-six inmate-on-inmate assaults during the period 1980-82. This is the same total as that advanced by defense witness Reynolds, although the figures for individual years vary between the two witnesses. The similarity of the figures and the fact that Fogel pointed out the inaccuracies of the Serious Incident Reports seriously undermines the magistrate’s finding of a “serious question” about the validity of Fogel’s figures.

Aside from this misreporting, there is also evidence of serious underreporting of violence. The magistrate acknowledged that “[ajssaults undoubtedly occur which do not appear in the records,” but he found that the lack of evidence specifically detailing the extent of underreporting made it impossible to consider these assaults. Sergeant Hotchkiss testified, however, that one-third to one-half of reportable incidents are never filed on Serious Incident Reports. According to him, the correctional staff at VSP “didn’t want you to see anything. They did not want you to be reporting anything____ Don’t bother me with this stuff.” Inmates Crute and Collins testified that guards often witnessed assaults and took no action; according to Crute, guards interceded in only about half of the assaults he has witnessed. Guard intervention appears to depend upon the proximity of back-up officers: Inmate Borkman testified that “if there is a fight where there is two or three guards, they react. If there is only one, it’s ignored.”

In addition to violence officially observed but unreported, the evidence makes clear that if violence is not officially observed, the inmates’ fear of being retaliated against for “snitching” severely limits any possibility that inmate witnesses will make any report. The magistrate found that among inmates there is a code of silence, enforced with threats of reprisal, which discourages reporting to officials. The practical effect of this code was illustrated by specific incidents. Correctional sergeant McCann, who investigated serious incidents at VSP, testified that inmate Robinson was stabbed to death in the mess hall during a movie because he was suspected of being a “snitch,” even though in fact he was not reporting to prison officials. Inmate Abbey testified that he was stabbed after reporting a theft from his cell. Inmates Carter and Doe (A) did not report their stabbing injuries or cooperate in prosecution of suspects because, as Carter testified, “if I had, I would have been labeled as a snitch and I could have been killed.”

The magistrate appeared to adopt the view of corrections expert Sigler, a defense witness, that the frequency of violent incidents at VSP is “[bjetter than some and worse than other[j” prisons. Sigler did not, however, support this general assessment with systematic statistical comparisons, and the statistics he did cite underscore the risk of violence at VSP. He referred to the state prison in Lucasville, Ohio, involved in Rhodes v. Chapman, supra, where an inmate population two and one half times the size of VSP’s suffered fewer inmate murders over the three-year period 1980-82. Corrections expert Fogel, an inmate witness, by contrast, explained in detail his statistical analysis and his conclusion that the rate of inmate-on-inmate assault with weapons was eight percent higher at VSP than at San Quentin in California, conceded by both sides to be one of the country’s most violent prisons. He labeled this a conservative estimate of the true level of violence at VSP. Fogel further testified that VSP’s rate was roughly equal to or slightly less than that at States-ville in Illinois, which he described as “a pretty rough place.” The magistrate discredited Fogel’s study because of the discrepancies between his figures and official figures. This appears to me to be unfounded because, as noted above, Fogel fully explained the discrepancies.

The brutality of specific incidents buttresses the statistical proof of frequency of violence in demonstrating a pervasive risk of harm. In discounting the stories of violence told by individual inmates, the magistrate referred generally to the testimony of six inmates who stated that they had never suffered an assault requiring medical attention. One of those inmates admitted, however, that he had been transferred out of VSP because prison officials were aware of threats against his life. Another of the six related an incident in which he was robbed at knifepoint in his own cell and then locked in with a padlock. More to the point, the fact that some inmates escape violence is clearly outweighed by the brutality suffered by those less fortunate. Inmate Puckett was padlocked into his cell and burned to death in a kerosene blaze. Inmate Wilkens was murdered on the ball field, stabbed with a knife and bludgeoned with a horseshoe peg. Inmate Robinson was stabbed to death in the mess hall during a movie. Inmate James had his face doused with acid in the mess hall, and three or four bystanders received severe burns. Inmate Stith was hit with acid while in his cell. Inmate Yeary suffered severe burns when he was padlocked into his cell and doused with buckets of scalding water heated by a “stinger,” a contraband electrical device with metal prongs that carry a charge. Inmates reported numerous other instances of being attacked or robbed in their own cells, usually at knifepoint. (Shrader: robbery); (Borkman: two robberies); (Nutall: apparent homosexual attack by four men); (Kendrick: robbery); (Stump: robbery). The common areas of the prison are no safer than the cells. (Page battered unconcious in the day room); (McCoy stabbed with an ice pick); (Abbey stabbed for reporting a theft); (Carter stabbed by mistake); (Stump stabbed in the course of an argument); (Doe (A) stabbed six times by two assailants). The situation at VSP may perhaps not be a reign of terror, but the entire record reveals substantially more than “isolated incidents” of violence, 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).

In finding the absence of a pervasive risk of harm from violence, the magistrate gave great weight to evidence showing that many violent incidents were related to drug trafficking and other illicit activities and that inmates who avoided these malefactions were “relatively safe.” He appears to have discounted the risk of violence as resulting in large part “from the voluntary association with an illicit activity.” It is true that many acts of violence stem from wrongdoing in prison. Corrections officials testified that inmates suffer violence “[b]y dealing in drugs, gambling, not paying off their debts.” E.E. Wright, Jr., chief of security at VSP, specifically cited inmate Borkman, a victim of at least two knife-point robberies because he is a “storeman,” bringing in goods such as cigarettes from the outside and reselling them to inmates at a profit. Inmate Borkman himself testified that “every act of violence is probably drug-related, whether it be for the money, the debts, or somehow it’s drug-related.” Former inmates John Doe B and John Doe C testified that inmates “get in trouble” because they “create their own problems” by stealing, “play[ing] slick games,” and cheating, rather than minding their own business.

I am unaware, however, of precedential support for the magistrate’s apparent conclusion that the inmates as a class have assumed the risk of violence because some engage in illicit dealings in prison. To the contrary — “equitable relief from unconstitutional living conditions should [not] be denied because the conditions were created in whole or in part by some of the prisoners.” Blake v. Hall, 668 F.2d 52, 58 (1st Cir.1981), cert. denied, 456 U.S. 983, 102 S.Ct. 2257, 72 L.Ed.2d 862 (1982); Ramos v. Lamm, 639 F.2d 559, 569 (10th Cir.1980), cert. denied, 450 U.S. 1041, 101 S.Ct. 1759, 68 L.Ed.2d 239 (1981). Moreover, the evidence demonstrates that numerous acts of violence are suffered by inmates who are innocent bystanders or have no stated connection to illegal activity. There was specific testimony that not every inmate who suffers an assault or a theft is involved in improper activity. Inmate Robinson was stabbed to death in the erroneous belief that he was an informer; inmate Abbey was stabbed for reporting a theft from his cell; inmate Carter was stabbed by an inmate who immediately apologized for getting the wrong man; three or four inmates, including Krueger and Lindh, suffered se vere acid burns when hit by stray acid aimed at inmate James. Of the many robberies and assaults mentioned in inmate testimony, most were never even suggested to be related to illicit activity: Nutall suffered an apparent homosexual attack; Kendrick lost a diamond ring in a robbery the day after his arrival at VSP; Puckett was burned to death in his cell. Even former inmate John Doe B, who stated that inmates are safe if they mind their own business, admitted that inmates are often assaulted and robbed because they possess items that others want and are perceived to be too weak to defend them. I am convinced by the frequency and brutality of violence at VSP that the inmates there suffer a pervasive risk of harm.

Among the group of concededly dangerous and violence-prone men confined to VSP, this risk of serious harm is exacerbated by the free availability of contraband weapons within the prison. Although the magistrate is surely correct that virtually any item, from ballpoint pens to broom handles, may be turned to use as a weapon, some weapons are more effective than others and their presence facilitates violence. A homemade knife, acid, gasoline, or kerosene, may embolden inmates who would be less aggressive wielding a ballpoint pen. Even more importantly, these lethal weapons inflict substantially more severe injuries.

The magistrate’s opinion contains no express findings on the prevalence of weapons at VSP, but merely the statement that “weapons do exist.” He apparently felt that further findings were obviated by his conclusion that the risk of violence from weapons was not “unnecessary,” because the risk comes “not ... from the weapons themselves ... [but] from the use of those weapons.” The majority has determined this finding to be partially inadequate and has remanded for the magistrate “to determine the extent of [the danger posed by weapons illicitly fabricated by inmates], the defendants’ response thereto, and whether this danger and response or lack of response rise to the level of an eighth amendment violation.” At 987. I am firmly convinced that the evidence demonstrates that contraband weapons are in plentiful supply at VSP, and that these weapons facilitate more frequent acts of violence and more severe injuries. I do not think a remand for further evidence is necessary to demonstrate this.

The defendants do not directly deny the existence of large numbers of weapons at VSP, so the testimony of the inmates’ witnesses was largely uncontradicted at trial. On the prevalence of homemade knives, incontestably the weapon of choice at VSP, inmate Yeary testified that there are “[q]uite a few of them. They are easy to get.” The knives are made in the shop, e.g., in the chair factory where Yeary worked. Sanding discs and grinding wheels are used to sharpen pieces of metal smuggled out of the metal shop: Yeary testified that “when you’re going to lunch, you come right by the metal shop. ... [T]he window is right there. They [conspiring inmates] will pass you out a piece of metal.” According to Yeary, one inmate made six knives in two days, selling them to other inmates, and twenty to thirty have been made in a couple weeks’ time. Corrections sergeant Hotchkiss testified that in two years, he alone (one of a staff of hundreds of corrections officers) has taken about seventy-five lethal knives from the cell houses. These are substantially tooled and machined and generally ten to twelve inches in length. He has also recovered about thirty-five new hacksaw blades, as well as many lengths of cold water pipe. He identified the number of weapons at VSP as being 250 percent higher than the number at a prison in Iowa where he worked prior to joining VSP. Inmate Vin-nedge testified that “the knives, of course, you can find them all over the institution. And, of course, the acid which has already been mentioned, and fire.” The frequency of serious assaults with weapons, however, is the best indication of the easy access to contraband weapons.

Drug use and drug trafficking are the handmaidens to free access to weapons in exacerbating violence at VSP. Although there is widely conflicting evidence on the incidence of drug abuse among VSP inmates, the magistrate found that “there is indisputably a significant problem.” Defense witnesses estimate that the number of inmates using drugs intravenously is ten percent or less of the prison population. Corrections sergeant Hotchkiss, testifying for the inmates, estimated that over half of the inmates are “shooting up, using the needle.” Inmate Borkman testified that “[o]ver fifty percent of the inmates are either on drugs or pushing drugs.” From 1979 to 1982 eleven cases of drug overdose occurred, one of which resulted in death. Several inmates testified that marijuana is readily available, as is an analgesic called Talwin, and that they regularly used both. Drugs are “all over the place” and, for inmates with money, “about as available as cigarettes,” according to the testimony of inmates Krueger, Evans, Freeman, and Abbey. Other drugs are present in substantially smaller quantities. Sergeant McCann testified that he had recovered Talwin in lots of 60,100, and 36 tablets in a one-year period, although the average amount found during cell shakedowns is ten tablets. Marijuana is found mainly in small packets, although up to one pound has been recovered at a time.

The magistrate found no constitutional violation on these facts, however, because he found that the only inmates harmed by the drug trade were those who voluntarily associated themselves with it and that the defendants had taken “some reasonable, albeit unsuccessful, measures to halt the introduction of drugs into the institution.” His statement that “there is no evidence that the use of Talwin or other drugs has significantly increased the incidents of larceny, robbery, etc.” conflicts with his finding at the same time that most or all violence is related to the drug trade and other illicit activity. In addition, it ignores testimony of witnesses for both sides that many acts of violence stem from wrongdoing in prison. Corrections officials testified that inmates suffer violence “[b]y dealing in drugs, gambling, not paying off their debts.” Inmate Borkman testified that “every act of violence is probably drug-related, whether it be for the money, the debts, or somehow it’s drug-related.” Reviewing all the evidence it is impossible not to conclude that the drug problem at VSP is significant and that it contributes importantly to the level of violence among inmates.

II. Significant Mental Pain

I can agree with the magistrate and the majority that “fear of constitutional dimensions [involves] more than simple anxiety. [An inmate] must demonstrate anxiety on a level such as would interfere to some degree with his everyday functions____ [T]here must be evidence of serious mental and emotional deterioration attributable to the fear of constant danger from assaults.” I cannot agree, however, that the inmates have not carried their burden under this standard, nor with the magistrate’s conclusions that “the evidence of any real fear of assaults at [VSP] is virtually non-existent” and that fear of assault “is transient, momentary, and does not cause any mental pain to the general population.”

It is true that defense witnesses who had visited the prison and talked with the inmates reported a lack of fear and tension in the prison, (corrections expert Sigler; corrections expert Hutto; psychology expert Samenow). VSP warden Booker and his administrative assistant Sheppard testified that at one meeting of the Inmate Advisory Committee Booker asked the inmate members if they were “scared or afraid to walk the yard” for fear of assault. The members replied that they were not afraid. Inmate Evans stated at that meeting that although he saw things he did not like, he was not afraid. Former VSP counselor Coble testified that of 2145 inmate grievances filed in 1982, only ten “related to violence,” and some of those were duplica-tive. The magistrate also noted that certain inmate witnesses admitted that they were not fearful, but he did not cite particular testimony. Finally, the magistrate appeared to give substantial weight to his own observations made during an unan nounced evening tour of VSP, during which he found “a striking lack of tension.”

My review of the entire record impresses me that the evidence relied upon by the magistrate suffers from inherent unreliability, particularly the testimony of prison visitors observing inmate behavior under plainly artificial conditions. Moreover, that evidence is clearly outweighed by other evidence to the effect that inmates have a great fear of violence that causes them significant mental pain and interferes substantially with their everyday functions.

The testimony of inmate Evans is revealing. In response to the question “Would you call yourself afraid?” he answered “no.” But he also testified that he was “acutely aware” of a sense of physical danger and that “[tjhere are times when the atmosphere is very tense, when people are obviously afraid____” His professed lack of fear is apparently one aspect of his attempt to “intimidate ... or show physical or emotional strength” and not show any form of weakness. Evans refuses to admit that he is afraid in order to maintain his “tough guy” image, a necessary defense against being victimized. At the same time, he has developed routines and methods of self-help designed to afford a measure of genuine security. He testified that he would change his “movement pattern” and “not go out in the population if I feel there is danger there.” In the mess hall he and a group of trusted inmates sit together in a group, facing each other and maintaining a lookout for surprise attack over each other’s backs.

This pattern of projecting a tough image and pursuing self-help recurs in the testimony of other inmates. Inmate Joseph, despite admitting that he didn’t “have any fears” walking around the prison, stated that he could “feel the tension” and coped. by not “associat[ingj myself with those things, with violence.”- Inmate McCoy testified that “if you don’t do something to make some kind of a statement about yourself generally you will be used in some type of way.” Corrections expert Fogel, a witness for the inmates, testified that inmates are reluctant to acknowledge violence or express fear, because that would connote weakness and might be considered informing. A casual attitude or macho image could be a mask for fear. Defense expert psychologist Samenow conceded that the inmates he interviewed who expressed no fear may have put up a tough front to avoid confessing weakness. Inmate McNair testified that “I walk in a manner that says I don’t care. But then again, when I walk, I’m always on my guard there because you never can tell what may happen from one minute to the next____ [I]f you don’t appear relaxed or carry yourself in a state where you don’t care, then that shows a certain amount of weakness there, and when you show weakness there, then you become pretty much prey to be victimized.”

Inmate testimony also clearly shows that all inmates practice various forms of self-help against violence, because, as Inmate Doe A testified, “[i]f you are not prepared to protect yourself, you’re not going to be protected there.” The inmates do not feel comfortable relying upon the guards for protection. Some forms of inmate self-help are relatively benign, for example the common practice of sitting in groups at meals to cover each other’s backs. (Borkman, Vinnedge, Lindh, Evans, McNair). Other forms contribute to the “serious mental and emotional deterioration” of the inmate practicing them. For example, Inmate Borkman keeps himself locked in his cell as “the best way to avoid conflict” and danger. He won’t go to “medical,” “dental,” the movies, the ball field, recreation, or “anyplace to expose myself to potential threats.” Inmate Krueger refuses to go to the mess hall during lunchtime. ■ Corrections sergeant Hotchkiss testified that inmates cut six-inch square windows in shower .curtains to look through and detect attack whil.e in the shoyver, and that inmates may keep their, possessions to a minimum to deter robbery. Most dangerous are those methods of self-help that further undermine the security in the institution. Chief among these áre the practice of carrying contraband weapons in self-de fense, (inmate Yeary: many inmates carry knives to protect themselves); (inmate Nu-tall has carried numb chucks, a martial arts weapon), and that of refusing to report or assist in the investigation of wrongdoing for fear of reprisal. (Carter, Doe (A)). In addition there is the common practice of inmates locking themselves into their own cells, using a padlock or, after prison officials banned the use of padlocks for this purpose, a sliding locking device fabricated in the prison shop.

Considerable testimony of subjective feelings of fear reinforce the objective evidence of substantial mental pain suffered by the inmates. Inmate Shrader testified that he feels “[hjighly paranoid, for the first time in my life.” According to inmate Krueger, the atmosphere in the prison is “[v]ery guarded, very stoic, very wary.” Inmate Vinnedge “feel[s] that security is lacking seriously, very seriously.” Inmates Evans, Joseph, and Freeman all testified to the high level of tension among inmates. The inmates’ corrections expert witness Fogel testified that interviews with inmates revealed to him that “[pjeople are running scared. Things are tense.” The inmates feel that “the thing you’ve got to do here is stay alive during the day____ [They] just want to make it through each day____” Criminology expert Korn, testifying for the inmates, recounted his observation based on conversations with inmates that “violence and the fear of violence, particularly the fear of violence, is univer-sal____ [Inmates are] pleading for greater security; in terror of fire bombing, in terror of being incinerated.”

On the record as a whole, I am firmly convinced that the magistrate clearly erred in failing to find the significant mental pain sufficient to comprise a constitutional violation.

III. Unnecessary and Wanton Infliction

The magistrate found that the generally high risk of violence at VSP was justified, even if it caused the inmates significant mental pain, because such risk is “a necessary adjunct to any sentence of confinement [in a maximum security institution].” Institutions like VSP, he noted, house violent men serving long sentences, and “[i]t is only natural to expect violent men to do violent deeds____ It follows, therefore, that a high risk of assaults in a maximum security institution is not totally without penological justification.” Moreover, he found that prison officials at VSP were not guilty of a willful and wanton failure to take reasonable security precautions to protect inmates, citing the absence of double celling, the relatively high guard-to-inmate ratio, the use of a classification system to separate known inmate enemies, routine inmate searches and cell shakedowns to discover contraband, and adequate lighting and guard deployment.

It is true that officials at VSP have implemented important security precautions to reduce the risk of violence. Despite these precautions, however, inmates continue to suffer violent assault and death. This is due in part to the fact that many existing security precautions are ineffectively implemented. For example, the favorable guard-to-inmate ratio must be discounted in light of the fact that corrections officers may work as many as forty hours of overtime per week. If an officer volunteers for overtime, he must work a minimum of twenty-four hours of overtime per week. Sergeant Hotchkiss testified that toward the end of the commonly-worked sixteen hour shifts, guards “couldn’t even protect themselves ... much less protect the inmates____” Substantial overtime is resorted to because there are only 307 security employees for 341 established security staff positions. The inmates’ corrections expert Keve testified that the pattern of guard deployment may be hypothetically “adequate when everything is going all right, ... if all those officers are there and not distracted, and there is no tension in the building____ But when anybody is pulled away, distracted, absent, it’s not enough. In other words, there is no margin of safety in it.”

The policy of inmate and cell searches is also ineffective in practice. Inmates Vin- nedge and Yeary testified that the patdown searches consist of quickly slapping an inmate’s sides and the outside of his legs and can easily be defeated by concealing contraband in the waistband over the groin. Sergeant Hotchkiss corroborated this, noting that “the pat-down system is just somewhat of a joke.” Guards “are just not going to” frisk an inmate’s groin, inmates wear many layers of heavy clothing to frustrate discovery, and some inmates are filthy with excrement which deters the guard’s search. More importantly, during mass movements between the cell houses and the mess hall large numbers of inmates must be frisked, leading to cursory searches to speed the process. The metal detectors currently in use are unhelpful because “half the time they don’t even work, plus in order to make it work you have to hold it approximately 4 inches from the individual at a maximum.” Former VSP assistant warden Reynolds conceded that an airport-style walk-through metal detector would enhance security, although defense corrections expert Hutto noted that personal searches would still be necessary to detect nonmetal weapons and drugs. Cell shakedowns are frequently ineffective, because of the pressure from supervisors to move along after only a cursory search, in order to cover more cells. Inmates testified to the ineffectiveness of searches and shakedowns: Inmate Evans, for example, noted that a shakedown of his cell failed to discover a bag of thirty syringes and a telescopic rifle sight. Inmate Vinnedge testified that, while searching his cell, guards spent fifteen minutes playing with a Rubik’s Cube and a Braille machine.

Inmate, Vinnedge testified that although a barbed; wire screen encloses the scrap metal pile, on occasion thé door to that area has been left wide open and accessible .without restraint to at least fifty inmates. The magistrate observed during his unannounced evening visit to the prison that “several of the screens on the back of the machine shop have been either pulled out or the bolts have been removed, and there is a space between, a larger space between the grid and the wall than there should have been.”

In addition to sometimes permitting the ineffective implementation of existing precautions, the trial record shows that prison officials have failed to address certain known security weaknesses. For example, despite what even the inmates’ expert described as “quite good classification procedures” on the state level for determining institutional assignment (Keve), there is no system at VSP for housing segregation according to inmate age, type of offense, length of sentence, or disciplinary record (Shrader; Borkman; Collins). The inmates’ psychiatric expert testified and the magistrate conceded the obvious, that passive, weak, and young inmates will be victimized if confined in close proximity with stronger and more aggressive inmates.

Defense witnesses conceded that there are numerous spots in the prison that are blind to the guards, most notably the stairwells in the cell houses. These locations are prime places for assaults and serve as staging areas for assaults committed elsewhere.

A glaring security weakness indicated by the evidence is the failure of prison officials to pursue some program of searching employees in order to deter them from smuggling contraband, into the institution. The evidence is clear that large amounts of contraband, principally drugs, are available in VSP, and present security precautions appear to eliminate the inmates and their visitors as possible sources. The magistrate found that

the evidence concerning the illegal usagé ahd dealing1 of . drugs at the Virginia State Penitentiary is a- matter of gravé concern. The evidence fairly welleliniinates the inmates and.their.visitprs as,a source ,o.f the controlled súbstahces dm fcering the institution. The conclusion is not attractive. The facts may indicate Violations of federal criminal laws, but they fall short of establishing constitutional violations.

I find quite disquieting the magistrate’s curious failure to state the obvious conclu sion that some prison employees must be involved in contraband trafficking and to explore responsive actions. The inmates’ corrections expert Fogel testified that VSP is “one of the few maximum [security] institutions” that has no program of searching employees to prevent traffic in contraband. Such a policy was in place at VSP at one time, involving random searches of employees upon entering and leaving the prison, but it either was terminated about four and one-half years ago or is no longer being enforced. Defense expert Hutto testified that routine searches of corrections officers is a “bad correctional practice” because it undermines the morale of the staff, but he admitted that unannounced random searches should be undertaken if there was reason to believe that prison employees were smuggling contraband. There have been at least two instances in which VSP employees have been caught smuggling contraband. Sergeant McCann testified that one employee was terminated for bringing in marijuana and the other for counterfeiting; the counterfeiter was also suspected of drug trafficking.

The magistrate conducted a protracted trial in a fair and impartial manner and was faced with a difficult chore in evaluating eleven volumes of testimony and hundreds of exhibits. I am convinced, however, that the inmates carried their burden of proving a violation of their eighth amendment right to be free of cruel and unusual punishment as that right is currently interpreted. In my view the inmates proved significant mental pain suffered as a result of conditions inflicted “willfully and wantonly” by prison officials, as that term is used in the eighth amendment context. The failure to implement effectively the current security precautions, as well as the failure to address known security weaknesses, is totally without penological justification and should be addressed by judicial order.

The inmates seek only a declaration that conditions of confinement at VSP fall below the eighth amendment minima and an injunction requiring the defendants to redress conditions to that level. This should involve minimal judicial oversight. Controversies involving constitutional claims against prison administrations more often than not create tension between the executive and judicial branches and between state and federal governments. The federal courts of course are not equipped and should not even attempt to run state prisons. We must at times, however, accept the difficult role of referee in the conflict between claims for basic human rights by society’s least admired individuals and claims for effective prison administration by beleagured state officials. Any effective means devised by prison officials to reduce the hazards of prison life to a constitutional level should satisfy any judicial order. See Gates v. Collier, 501 F.2d 1291, 1310 (5th Cir.1974). The details of the appropriate remedial measures should, of course, be considered initially by a magistrate. The record strongly suggests, however, some measures that should be considered. Plainly VSP should follow the example of most other maximum security prisons and revive its prior policy of searching employees to deter their participation in contraband trafficking. In addition, it would seem a simple matter to base inmate housing assignments upon some classification scheme, particularly as the information required to execute such a scheme is already compiled by the state and forwarded to VSP. Similarly, improvements in reporting, recordkeeping, and search procedures could be accomplished by a minimally burdensome program of retraining of corrections officers. To control the most important security problem, the prevalence of contraband weapons, stricter controls on scrap metal and flammable liquids and acids could be instituted, in the form of more secure storage and more restricted access. To prevent fabrication of weapons from these raw materials, inmate activity in the prison shops could be more closely supervised. Although some control of contraband may be obtained by the installation of equipment such as walk-through metal detectors, it is apparent that any significant increment to security at VSP can be achieved only by the hiring of additional corrections officers, both to expand the number of security posts and to reduce the heavy overtime burden which must impede even the most conscientious officer. If, for some reason imperceptible from this distant vantage, these steps are unworkable, then other measures capable of achieving similar results could be undertaken.

In determining the constitutionality of conditions at any prison, be it a minimum or maximum security facility, we cannot view the individual conditions challenged as distinct and unrelated, but instead must consider them as parts of a whole. Each condition affects the other — weapons facilitate and increase the degree of harm done in assaults, drug use and trafficking stimulate assaults, and inadequate security not only facilitates the abuse of the weak by the strong but engenders fear and isolation among all inmates. The evidence establishing the prevalence of weapons and drugs, the frequency of assault, and the inadequacy of security precautions add and give credibility to the express testimony of numerous inmates that they live constantly in fear, requiring them to pursue various methods of self-help ranging from relatively benign techniques such as sitting in groups at meals to safeguard each other from ambush, through individually deleterious methods such as withdrawal from the life of the prison into isolation, to socially dangerous methods such as carrying weapons for defense. There can be no doubt of the awareness of prison officials of the existence of such fear.

In the past five years the inmate population at VSP has suffered seven murders and many reported serious assaults, and all parties concede that numerous acts of violence go unreported because the victims fear reprisal from the culprits. Without question, a determined inmate can manufacture and possess dangerous weapons, and obviously the more hardened, violent, and crime-sophisticated inmates are the ones who pursue most assiduously this illicit activity. Any judge can certainly appreciate the magistrate’s expressed concern that in dealing with such dangerous criminals “[sjociety, too, has its rights,” and I readily agree that the state must be allowed greater latitude in structuring the institutional treatment of these men. But I cannot agree that the deterrent effect of penal confinement may constitutionally be augmented by permitting the inmate-on-inmate violence that my review of the whole evidence firmly convinces me exists at VSP — regardless of the fact that the prison population includes the most hardened of criminals. Wycoff v. Brewer, 572 F.2d 1260, 1267 (8th Cir.1978). “[Pjart of the penalty that criminal offenders must pay for their offenses against society,” Rhodes v. Chapman, 452 U.S. 337, 347, 101 S.Ct. 2392, 69 L.Ed.2d 59 (1981), cannot include the very great risk of injury from violence inflicted with contraband weapons and in incidents related to the use and dealing of contraband drugs, when the prevalence of both is the result of willfully lax and ineffectively executed security precautions. Permitting this violence to continue at VSP on the ground that fear among inmates is penologically justified abdicates to inmates control over the conditions of confinement. Blake v. Hall, 668 F.2d 52, 58 (1st Cir. 1981), cert. denied, 456 U.S. 983, 102 S.Ct. 2257 (1982). This necessarily means control by the cruelest and most violent. It may well be that the eighth amendment could have been interpreted to apply to inmates in different degrees according to their criminal personalities, but it has not been so interpreted.

Unless we are to hold that a prison population is not entitled to the protection of the eighth amendment when the inmates themselves inflict some of the suffering, I feel that we should grant to the inmate plaintiffs an injunction requiring the defendants to bring conditions at VSP to the minimum level required by the constitution.