From Negombo to Mahara, Magazine Prison, Kuruwita and Pallansena, the question is no longer simply why prisoners are rioting. It is why a system repeatedly placed on high alert appears unable to prevent the next eruption.
Sri Lanka has a habit of discovering the condition of its prisons only when they become violent. Overcrowding, prolonged remand detention, narcotics networks and ageing institutions can exist for years as chronic problems, serious enough to acknowledge but apparently capable of being managed indefinitely. Then a prison erupts, people die,reinforcements arrive and the vocabulary changes abruptly to security and emergency. Order is eventually restored, the immediate crisis passes and attention shifts elsewhere. What has happened over the past month is more difficult to explain away. Negombo was followed by Mahara; Mahara by violence at Colombo’s Magazine Prison and Kuruwita; and on the same day as the Kuruwita deaths, tear gas was used against protesting inmates at Pallansena. The disturbing question is no longer simply why these prisons became violent. It is why, after every incident supposedly places the system on alert, another institution appears capable of erupting.
That distinction matters. Sri Lanka has shown that it can regain control once disorder has begun. Police special forces, the military and navy were brought in after the catastrophic violence at Negombo. When Mahara erupted on 1 August, police, military personnel, drones and a helicopter were deployed, and authorities announced that security was being strengthened across major prisons. Yet within days, about 40 prisoners at Magazine Prison reportedly breached their cells and fought, leaving one dead; two more prisoners were killed in violence at Kuruwita, while officials used tear gas to control unrest at Pallansena. The incidents need not have been centrally coordinated for the sequence itself to be significant. Tactical control of one prison is not the same thing as strategic control of a prison system. If additional security follows one riot but cannot prevent serious disturbances elsewhere days later, the weakness may lie less in the ability to end disorder than in the capacity to anticipate it.
The individual incidents should not be collapsed into a conspiracy before evidence exists. Drugs, rival criminal groups, personal disputes, transfers and local security failures may have played different roles in different institutions. But insisting upon those differences should not become an excuse for refusing to see a pattern. Systems do not fail only when every institution fails in exactly the same way. They fail when the same vulnerabilities repeatedly produce different crises. What Sri Lanka now needs to understand is why tensions that should be detectable inside closely controlled institutions are repeatedly reaching the point of open violence.
One explanation lies in numbers so extreme that the word “overcrowding” no longer adequately conveys them. More than 39,000 inmates are being held in a prison estate designed for roughly 10,000 people, according to figures cited this week. At Mahara alone,around 4,100 prisoners were being held at approximately four times capacity when violence broke out. These are not simply uncomfortable prisons. They are institutions being asked to exercise precise control over populations radically larger than those for which their buildings, staffing structures and systems of classification were designed. That has direct consequences for security. Prison management depends upon classification: knowing who presents a threat, who belongs to which criminal network, which groups must be kept apart, who may be vulnerable and where tensions are developing. It requires officers to know the social geography of the institution almost as closely as its physical one. But classification becomes vastly harder when a facility houses several times its intended population. Rival prisoners are more difficult to separate, cells and communal areas become congested, movement becomes harder to supervise and officers are required to monitor far more relationships than an adequately staffed system should demand of them. In such an environment, restoring control after violence is possible; maintaining control before violence becomes substantially harder.
This may help explain why Sri Lanka appears to be fighting prison crises sequentially rather than systemically. After every disturbance, resources can be concentrated on the institution that has just erupted. Additional officers can be sent there, inmates confined, cells searched and security tightened. But if the underlying intelligence picture is weak across the prison estate, pressure does not disappear simply because one facility has been secured. Rivalries, narcotics networks and grievances remain elsewhere. Prisoners are moved between institutions. Criminal relationships cross prison boundaries. What looks like a local security problem may therefore be part of a much wider network of risk that cannot be contained by treating each prison as an island.
The transfers that followed Negombo illustrate why system-wide management matters, although there is no evidence that they caused the later riots. Prisoners were dispersed to other institutions after the July violence, and the Human Rights Commission subsequently began investigating not only what happened at Negombo but events involving transferred inmates elsewhere, including allegations of ill-treatment and deaths in custody. It also reported that its officers were initially denied access to Welikada because prison officials said there was an “unstable security situation”. Transfers are sometimes necessary after a major disturbance, but they also demonstrate a fundamental fact about a national prison system: decisions made in one institution can alter the security pressures inside another.
There is an equally serious question about organised crime. Authorities have linked the latest Magazine and Kuruwita violence to narcotics dealers and rival criminal gangs. If that assessment is correct, simply identifying drugs as the explanation would be inadequate. Prisons exist precisely because the state expects them to contain dangerous people and disrupt criminal activity. When rival networks retain the ability to exercise power inside custody, communicate, acquire contraband and mobilise violence, imprisonment has changed their location without necessarily breaking their organisation.
Every packet of narcotics, illicit phone or weapon found inside a prison also tells a story about the institution around it. Contraband does not pass through concrete by itself. It crosses a perimeter, moves through a controlled environment and reaches somebody who is not supposed to possess it. If illicit economies become sufficiently entrenched to sustain competing criminal groups, the relevant question is not only which inmate possessed the contraband but how the chain delivering it was allowed to exist. Security based predominantly on reacting to inmates after rules have been breached will always arrive too late. Effective prison control depends upon intelligence, surveillance, staff integrity, classification and the ability to identify networks before their rivalries turn a cell block into a battleground.
The pressure entering this system is extraordinary. Official prison statistics show that narcotics offences accounted for about two-thirds of direct admissions of convicted prisoners in 2024. Sri Lanka also maintains a large remand population – people held while criminal proceedings remain unresolved rather than serving final sentences. These are not separate facts from the riots. They help explain the conditions in which the prison service is attempting to maintain control. A delayed case becomes another remand prisoner. A drug policy heavily dependent on incarceration becomes another person added to an already saturated institution. Weak alternatives to custody eventually become a prison-management problem.
The result is a system in which the most dangerous prisoners can become harder, rather than easier, to control because scarce custodial resources are being spread across such enormous numbers. That is why reducing unnecessary incarceration is not the opposite of taking prison security seriously. It can be essential to it. A correctional service with finite officers, cells and intelligence resources should be able to concentrate them on the people who actually require high-security detention. Violent offenders and organised criminals should be controlled rigorously; that becomes harder when the institution managing them is simultaneously carrying thousands of people who might safely be dealt with through bail, treatment, probation, rehabilitation or community supervision.
There is another possible feature of repeated unrest that deserves investigation: contagion. Prisons are closed institutions, but they are not information vacuums. Prisoners know what occurs elsewhere through families, lawyers, illicit communications, transfers and institutional networks. A major disturbance in one prison can alter expectations in another, whether by encouraging imitation, intensifying grievances or prompting security measures that themselves create new tensions. There is not yet evidence to conclude that such a dynamic explains Sri Lanka’s current sequence. But after unrest has moved across several institutions, it would be negligent not to examine whether one crisis is affecting the conditions in which the next takes place.
That is ultimately why a succession of separate inquiries will not be enough. Sri Lanka needs to know not merely who started the violence at Magazine Prison or Kuruwita, but whether officers elsewhere had received warnings; whether rival groups had been identified in advance; whether prison intelligence is shared nationally; whether transfers altered gang balances; whether contraband routes operate across multiple institutions; and whether the same criminal networks are appearing repeatedly. A prison-by-prison investigation can establish culpability. Only a system-wide investigation can establish why, despite heightened security, the next prison remains vulnerable.
None of this excuses prisoners who kill, assault officers or organise criminal activity. A functioning justice system should be capable of recognising individual responsibility without pretending that institutions have none. Indeed, prisons are built on the assumption that some people inside them will be violent, organised and resistant to authority. If the explanation for a prison losing control is simply that dangerous criminals were present, then the explanation has inadvertently identified the very task the institution failed to perform.
Sri Lanka has confronted deadly prison violence before. That history makes the present sequence more troubling because the country cannot plausibly describe the weaknesses now being exposed as unforeseen. The lesson of Negombo should not merely have been that Negombo required more security. The lesson of Mahara could not simply be that Mahara required it too. If Magazine Prison, Kuruwita and Pallansena can follow after security has already been tightened nationally, the country is confronting something more serious than a collection of unruly institutions.
It is confronting the limits of a model of control that becomes most visible only after control has already been lost.
The distinction is crucial. Sri Lanka can send additional forces to a prison and restore order. It has done so repeatedly. But the true measure of a functioning correctional system is not how efficiently it suppresses the riot that has already begun. It is whether it possesses the intelligence, capacity and institutional authority to stop the next one from beginning.
That is the question the sequence from Negombo to Mahara, Magazine Prison, Kuruwita and Pallansena now demands.
The most troubling fact is not that Sri Lanka has failed to regain control of its prisons. It is that, each time it does, another prison appears to lose it.
