What is Stockholm Syndrome

A hostage pleads with police not to hurt the man holding a gun to her colleague’s head. A kidnapped heiress is later photographed robbing a bank beside her captors. A woman in an abusive marriage lies to the neighbors, smooths things over, and says “he’s not a bad man.” Outsiders watch these scenes and ask the same baffled question: why on earth are they protecting the person who’s hurting them?

The popular answer is a single phrase, Stockholm syndrome, and it gets used for everything from kidnappings to bad boyfriends to a character in a streaming drama who grows fond of her jailer. The phrase is tidy. The reality isn’t. Behind it lies a 1973 bank siege, a psychiatrist who never interviewed the hostages, a handful of FBI studies, and a decades-long argument over whether the term describes a real condition or a story we like telling.

One point should be said bluntly at the start: it is not an official diagnosis.

So what’s actually going on?

This article answers what Stockholm syndrome is, where it came from, why captives sometimes bond with captors, and why many psychologists now prefer other explanations.

What is Stockholm syndrome, in plain language?

Stockholm syndrome is a label for a psychological response in which a person who is held captive, threatened, or abused develops positive feelings toward the one controlling them. The captive may defend that person, resist rescue, or distrust the people trying to help. The term describes an observed pattern of behavior; it isn’t a medical diagnosis and it appears in neither the DSM nor the ICD.

That’s the short definition, and it hides a few traps. “Positive feelings” can mean actual affection, but just as often it means compliance, gratitude, or appeasement that looks like affection from the outside. And “bonding” doesn’t require a long confinement; the original case lasted six days.

Researchers and law enforcement have described several behaviors that tend to be grouped under the label:

• Captives feel sympathy or loyalty toward the person holding them.
• They show hostility or distrust toward police, rescuers, or family members trying to intervene.
• They treat small kindnesses, such as food, a blanket, or a calm word, as enormously meaningful.

Some versions of the description add that the captor, in turn, may soften toward the captive. That’s sometimes called the reverse pattern, though it’s far less studied.

It helps to say what the term doesn’t mean. It doesn’t mean the person is weak, gullible, or stupid. It doesn’t mean they “chose” to stay in a free-and-clear sense. And it doesn’t mean they’re mentally ill. Most experts who study it describe it as an understandable, sometimes protective response to extreme danger.

Think of it like flinching. If a ball flies at your face, you flinch before you decide to. Bonding with a dangerous person who controls your survival can work the same way: automatic, fast, and not particularly logical. Observers who didn’t face the threat tend to judge the reflex harshly.

A word about usage online. Searches for “Stockholm syndrome” often lump together hostage cases, abusive relationships, cults, trafficking, and toxic workplaces. Those situations share some dynamics, yet they differ in danger, duration, and evidence. The more loosely the term gets stretched, the less it explains.

And that stretching is where most of the confusion starts.

What really happened at the Stockholm bank in 1973?

On August 23, 1973, a convicted criminal named Jan-Erik Olsson walked into the Kreditbanken on Norrmalmstorg Square in central Stockholm, fired a submachine gun at the ceiling, and took four bank employees hostage. He demanded money, a getaway car, and the release of Clark Olofsson, an inmate he’d befriended. Police brought Olofsson in, and the standoff stretched across about six days, with the hostages held in a vault.

As the days passed, something unexpected happened. The hostages appeared to develop positive feelings toward their captors, and some pleaded with police not to harm them. One hostage reportedly said she feared the police more than Olsson, since a shootout would most likely get the hostages killed.

Swedish psychiatrist and criminologist Nils Bejerot, who advised police during the crisis, coined the label “Norrmalmstorg syndrome” for what he saw. The name spread internationally as Stockholm syndrome. Notably, Bejerot never interviewed the hostages, which means the first and most famous description of the syndrome came from an outsider watching the news and the negotiations.

Forensic psychologist Amy Grubb made the point in a 2026 article in The Conversation that this origin story matters: a term built on observation from afar later came to be treated as a diagnosis. She argued that the response is better read as a survival strategy, and noted that some hostage behaviors resemble techniques used deliberately by police negotiators, who try to build a relationship with the hostage taker to gain influence over their behavior.

Consider what the hostages knew in that vault. They knew a gunman held their lives in his hands. They knew police were outside, could use gas or force, and might misjudge. They also knew that the man inside could, in any moment, decide to be kind or deadly. In that setting, “he’s human; he might not kill me” is a strategy as much as a feeling.

The aftermath adds a twist. Several former hostages later said they’d been trying to stay alive, and objected to being described as irrational or emotionally captured. Their accounts differ in the details, and memory after a traumatic event isn’t a perfect record, but the picture they paint is of people doing arithmetic under pressure, not people succumbing to a spell.

Here’s a detail worth sitting with. The most famous case behind the term involved hostages who were largely unharmed by their captors, which is itself part of why the dynamic emerged. Kindness from someone who could kill you lands differently than kindness from anyone else.

Is Stockholm syndrome a real psychiatric diagnosis?

No. Myth: it’s a recognized mental disorder with standard criteria. Fact: no major diagnostic manual includes it, and there’s no agreed definition or validated test, which is why many researchers consider it a descriptive label rather than a condition.

A well-known review by Mohammed Namnyak and colleagues, published in 2008 in Acta Psychiatrica Scandinavica under the pointed title “‘Stockholm syndrome’: Psychiatric diagnosis or urban myth?”, examined the literature and found it thin and inconsistent. The authors reported a lack of consensus on what the term means, a heavy reliance on media reports and anecdote rather than systematic study, and wide variation in what counts as a case. They questioned whether the phenomenon deserved to be treated as a distinct clinical entity.

That doesn’t mean the behavior never happens. Plenty of captives, abuse survivors, and others have described becoming attached to or protective of a person who harmed them. What’s disputed is whether it forms a coherent syndrome, how often it appears, and whether “syndrome” is the right frame.

Clinicians typically work from established diagnoses. A person who’s been held captive or abused might meet criteria for post-traumatic stress disorder (PTSD), complex PTSD (recognized in the ICD-11), depression, or an anxiety disorder, and those can be treated. Attachment to an abuser can be part of the clinical picture without being its own diagnosis.

The table shows something useful: the rows with the strongest clinical standing describe suffering and symptoms, while the more famous label describes a relationship pattern.

Why does the distinction matter beyond semantics? Because a diagnosis shapes treatment, insurance, and how courts and the public regard someone. A label with no clear criteria can be used loosely, sometimes to dismiss survivors (“she’s just got Stockholm syndrome”) and sometimes to explain behavior no one has bothered to examine.

Honest summary: real behavior, shaky concept.

What conditions make captives bond with captors?

Take Priya (an illustrative composite), kept in a locked apartment by a man who controls her phone, food, and money. He’s volatile, but on some days he brings her tea and says he’s sorry. By the third week she’s listening for his footsteps, tracking his mood, and feeling a rush of relief at a kind word. Her brain is doing what brains do: studying the person on whom survival depends.

FBI behavioral specialist Thomas Strentz described conditions under which this pattern tends to appear in a 1980 paper on law enforcement policy and hostage behavior. In summary, he pointed to four ingredients that are commonly cited:

1. The captive perceives a threat to survival and believes the captor is willing to act on it.
2. The captive perceives small kindnesses from the captor, in the context of that threat.
3. The captive is isolated from outside perspectives, and sees only the captor’s point of view.
4. The captive believes escape is impossible, or at least extremely dangerous.

These conditions don’t guarantee bonding, and they aren’t a validated prediction tool. They’re a description drawn from case experience in law enforcement. Still, they map well onto accounts from survivors of kidnapping, hostage-taking, and coercive abuse.

Notice the role of contrast. Kindness matters more when it follows fear. A glass of water from a stranger doesn’t register much; a glass of water from a person who just threatened you can feel like grace. Psychologically, relief is a powerful bonding agent.

Strentz also wrote about how hostages can help themselves, including by staying calm, avoiding confrontation, and building a human connection with the captor, which makes the captor less likely to see them as an object. That advice treats rapport-building as a deliberate survival tactic, which complicates the idea that the bond is purely irrational.

Duration probably matters too, as does intensity. Hours of captivity differ from months. A person held in a locked room differs from a person in a controlling relationship who has a job, a phone, and neighbors. Real life rarely supplies all four conditions neatly, which is partly why people disagree about when the label applies.

What’s most important is the direction of reasoning. These conditions describe the situation, not a flaw in the person. Put almost anyone in them, and you’d see shifts in thinking and feeling.

Why would anyone side with the person hurting them?

Because, from inside the situation, it can be the safest move available. Several overlapping explanations have been proposed, none of them complete, and each emphasizes survival.

The first is appeasement. If pleasing the captor reduces the odds of violence, then acting warm, loyal, or compliant is a rational response. Amy Grubb and others have suggested replacing the term Stockholm syndrome with something like appeasement, since it highlights function rather than pathology. Survivors of abusive relationships often say the same thing: “I learned exactly what kept him calm.”

The second is identification with the aggressor, a concept from psychoanalytic theory developed by Anna Freud in her 1936 book The Ego and the Mechanisms of Defence. She described how a frightened person may take on the traits or perspective of the one who threatens them, as a way of mastering fear. It’s a classic psychoanalytic idea, influential but hard to test experimentally, so I’d treat it as a lens rather than a proven mechanism.

The third is learned helplessness, associated with Martin E. P. Seligman, whose research (summarized in his 1975 book Helplessness) found that animals exposed to uncontrollable stress later failed to try escaping even when escape became possible. Applied to captives, the idea is that repeated experiences of having no control can dull the urge to resist. Later reinterpretations of Seligman’s findings emphasize that passivity may be the default and that learning control is what changes things; human cases are more complicated than the lab.

The fourth is simple dependence. When one person controls your food, safety, and information, you pay close attention to them. The attention can feel like closeness.

And the fifth is cognitive consistency. If you can’t leave, it’s easier on the mind to find reasons you wouldn’t want to: “He’s not so bad; he has a hard life.” That’s a way of making an impossible situation bearable.

Picture Daniel (an illustrative composite), held for hours during a robbery. The robber tells him, “Don’t worry, nobody’s getting hurt if you stay calm,” and offers him water. Daniel hears himself saying thank you. Later, he’s ashamed of that. But what else would a person do with a gun in the room?

These pieces fit together, and they share a message: the behavior makes sense. What looks irrational from the outside is often exactly what a smart nervous system does when it’s cornered.

How common is Stockholm syndrome, really?

Less common than the pop-culture version suggests. An often-cited figure from the FBI’s hostage barricade database, discussed by FBI crisis negotiator G. Dwayne Fuselier in a 1999 article, put the rate at roughly 8 percent of hostage victims showing signs associated with the syndrome. That number gets repeated everywhere; treat it as a rough indication, not a precise prevalence, since definitions and data collection vary.

Fuselier’s broader point was a corrective for negotiators. Hostages often don’t develop bonds with captors, and some develop the opposite: intense fear and hatred. Treating the syndrome as the typical outcome can mislead police into misreading a victim’s behavior.

Why is the real rate so hard to pin down? A few reasons stand out:

• There are no agreed criteria, so different studies count different things.
• Hostage situations are rare and chaotic, making systematic data collection difficult.
• Reports often rely on police observation or media accounts, not structured interviews with survivors.
• Victims may perform compliance for safety without feeling any attachment, which looks the same from a distance.

That last point deserves emphasis. If a hostage smiles at her captor to keep him calm, an observer may record “bonding” when the person was actually doing the cognitive work of survival. Behavior and feeling are different things.

Some studies of other populations, such as people in abusive relationships, find that attachment to abusers is common, but they use their own measures and populations, and they rarely call it Stockholm syndrome in the strict hostage sense. Applying a number from one setting to another is how statistics get stretched beyond their warranty.

So the honest answer is: nobody knows precisely, but the experience isn’t the default, and it’s rarer in hostage situations than television implies. Most hostages, most of the time, want out and want their captors held to account.

A statistic repeated for decades without a clear source should make any editor squint. This one has a source; the squint remains.

Does it happen outside hostage situations?

Possibly, though researchers prefer more precise terms for those settings. People have applied the idea of Stockholm syndrome to abusive relationships, coercive cults, human trafficking, and childhood abuse. The shared thread is a power imbalance with unpredictable harm and kindness, which is what most trauma bonding research is about.

Psychologist Dee L. R. Graham was one of the best-known voices extending the concept. In her 1994 book Loving to Survive, written with Edna Rawlings and Roberta Rigsby, she argued that the survival logic seen in hostage cases also operates in battered women and other groups subject to sustained coercion, and framed the response as a normal reaction to terror and dependence. Her work is influential in feminist psychology, though critics argue that stretching the label across so many contexts reduces its precision.

Psychiatrist Judith Herman, in Trauma and Recovery, described how prolonged trauma in situations of captivity, from domestic violence to political imprisonment, can reshape a person’s relationship to the perpetrator, who may become the main source of both fear and comfort. Her account emphasizes coercive control: isolation, unpredictability, humiliation, and enforced dependence. That’s the same toolkit used in many abusive households and organizations.

Imagine Lila (an illustrative composite), whose partner monitors her phone, belittles her friends, then showers her with gifts after outbursts. She tells her sister, “You don’t know him like I do.” What looks like loyalty may be a mix of fear, hope, financial dependence, and plain exhaustion. Is it Stockholm syndrome? Maybe the label doesn’t matter; the danger does.

Differences between settings matter too. A hostage situation usually ends in days, with police involved; an abusive relationship can last decades, with a legal and social life continuing around it. In the latter, ties like shared children, housing, money, and religion are anchors that no hostage faces in quite the same way.

The evidence for applying the syndrome to these contexts is mostly clinical observation, survivor testimony, and theoretical work, not controlled studies. That’s valuable, but it’s different from showing that one syndrome explains all of them.

In a related post on our site we look at how this plays out in jobs where a manager alternates cruelty and charm; that setting has its own research under the name abusive supervision. The core lesson transfers: when your safety or livelihood depends on someone, attachment becomes a survival tool.

What does it look and feel like from inside versus outside?

From outside, it looks like defensiveness. From inside, it feels like tangled loyalty, fear, relief, and doubt, often in the same minute. That gap is why friends and family so often make things worse by arguing the person out of their feelings.

Signs commonly described by survivors and clinicians include:

• Making excuses for the harmful person (“he’s under stress,” “she had a rough childhood”).
• Feeling grateful for small kindnesses and clinging to memories of good moments.
• Becoming hypervigilant to the other person’s moods, adjusting behavior to avoid a blowup.
• Pulling away from outsiders who criticize the person or suggest leaving.
• Minimizing or hiding harm from others, including doctors or police.
• Feeling guilt or disloyalty when imagining leaving, even when the situation is clearly harmful.

Inside, many survivors describe a sense of confusion and shame later, once they’re out. “How could I have loved him?” is a common question. The answer is in the circumstances: if you’re dependent on someone and can’t flee, your mind makes space for the person you must live with.

Consider Marcus (an illustrative composite), whose older cousin controlled him during a long, coercive stretch of his adolescence. Years later, he still sends the cousin a holiday message. His therapist doesn’t tell him he’s foolish. She asks what that message does for him, and what he’d lose by not sending it.

That question, “what function does this serve?”, is the heart of modern thinking about the response. It replaces judgment with curiosity.

For friends and relatives, a few cautions help. Avoid ultimatums, since they often push people back toward the harmful person. Avoid trashing the abuser, which can trigger loyalty and defensiveness. And avoid asking “why don’t you just leave?” as if the answer were obvious. Staying has reasons. Leaving carries real danger; research on abuse suggests that the period around separation can be among the riskiest.

What helps? Steady, nonjudgmental contact, practical support, and patience. The relationship you keep open might be the lifeline later.

How does the label itself hurt survivors?

It can make a rational survival response look like a defect. When police, courts, doctors, or the media talk about “Stockholm syndrome,” they sometimes imply the survivor’s judgment is distorted, which can undercut credibility or distract from the perpetrator’s actions. Amy Grubb warned that the label can make survivors feel their reactions are abnormal and shifts attention away from the circumstances that produced the behavior.

The most famous example is Patty Hearst, the newspaper heiress kidnapped in 1974 by a radical group. Weeks later she took part in a bank robbery with her captors, and was photographed doing it. Her defense argued she’d been coerced and brutalized, and commentators reached for the “Stockholm syndrome” explanation; she was nonetheless convicted in 1976. Her sentence was commuted in 1979, and she received a presidential pardon in 2001. Whatever one believes about her case, it shows how the label can be used as a story about someone instead of an investigation of what happened to her.

There’s a second problem: overuse. When a term gets applied to every troubled relationship, it blurs the specifics. A person in a situation of ordinary conflict isn’t in a hostage situation. A person who’s attached to a partner who sometimes behaves badly isn’t automatically “trauma bonded.” Over-labeling can pathologize normal complexity and trivialize genuine captivity.

A third problem is dismissal. Calling someone “Stockholm syndrome” can become a way of shutting down their account: “Don’t listen to her; she’s brainwashed.” That’s rarely helpful, and sometimes it protects the person who caused harm.

Language that centers function and circumstances does better. Instead of “she has Stockholm syndrome,” consider “she was in a situation where appeasing him was the safest option.” Instead of “he’s bonded to his captor,” try “he adapted to survive.” The facts are the same, and the framing respects the person.

For clinicians, the recommendation is to assess for trauma symptoms, safety, and coercive control, not for a syndrome with no criteria. For journalists, it’s to avoid using the term as a shorthand for victims’ behavior without explaining the context.

A small test for any writer: if you removed the label, would your description still make sense? If it does, you didn’t need it.

How can you help someone, or yourself, recover?

Safety comes first, and then the slower work of making sense of what happened. Recovery from captivity or coercive control typically involves addressing trauma symptoms, rebuilding connections, and loosening the grip of the bond, which tends to fade with time and distance, though not always smoothly.

Judith Herman’s model of trauma recovery, which has influenced clinical practice for decades, proposes three stages: safety, then remembrance and mourning, then reconnection. Her framework fits situations of prolonged coercion because it puts safety ahead of insight. Trying to “process” feelings while still in danger isn’t therapy; it’s risk.

Practical steps for someone recovering, or supporting someone who is:

1. Secure physical safety. If there’s immediate danger, contact emergency services. A safety plan, ideally made with a trained advocate or clinician, matters more than any conversation about feelings.
2. Keep the door open. Stay in contact without pressure. Isolation is a tool of coercive control, and connection undoes it.
3. Validate before you advise. “It makes sense that you felt that way” works better than “how could you.”
4. Get trauma-informed treatment. Approaches such as trauma-focused cognitive behavioral therapy and EMDR have evidence for PTSD; therapists who understand coercive control can help with the attachment piece.
5. Expect mixed feelings. Missing the person, feeling guilty, or doubting your memory are common and don’t mean you made the wrong choice.

Recovery usually isn’t linear. People can drift back toward a harmful relationship more than once before leaving for good, and that isn’t failure; it’s often how leaving works. Studies of attachment to abusive partners after separation suggest it tends to weaken over months, though it doesn’t disappear on a schedule.

If you’re struggling with thoughts of harming yourself, please reach out now. In the US, call or text 988; in Canada, call or text 9-8-8; in the UK or Ireland, Samaritans is available at 116 123 (Ireland also has Pieta at 1800 247 247); in Australia, Lifeline is 13 11 14; in New Zealand, call or text 1737. In an emergency, dial 911 in the US or Canada, 999 in the UK, 112 or 999 in Ireland, 000 in Australia, or 111 in New Zealand.

And if you’re the friend on the outside, keep this in mind: you don’t have to understand the bond to stay beside the person in it.

FAQs about What Stockholm Syndrome Is

What does Stockholm syndrome mean?

It’s a name for a pattern in which a captive or abused person develops positive feelings toward the one who holds power over them, such as defending them, siding with them, or distrusting rescuers. The term comes from a 1973 bank robbery in Stockholm, Sweden, where hostages seemed to side with their captors during a roughly six-day standoff. Despite its fame, it isn’t an official diagnosis, and researchers disagree about how common it is or whether it forms a distinct syndrome. Many now see it as a survival response, a way of staying safe with someone dangerous.

Is Stockholm syndrome a real mental illness?

No. It isn’t listed in the DSM or the ICD, and there are no agreed criteria for diagnosing it. A 2008 review by Namnyak and colleagues questioned whether it should be considered a psychiatric diagnosis at all, noting inconsistent definitions and reliance on media accounts. That doesn’t make the underlying experiences fictional. People who’ve been held captive or abused can have real symptoms, including PTSD, depression, and anxiety, and those are treatable. The behavior called Stockholm syndrome is better thought of as an understandable reaction to extreme circumstances, not a disorder in itself.

Why does it happen?

Several explanations overlap. Pleasing or calming a dangerous person can reduce the risk of harm, so appeasement works as a survival tactic. Isolation limits other perspectives, and small kindnesses feel huge when you’re terrified. Psychoanalytic theory suggests identifying with the aggressor reduces fear, and learned helplessness research suggests repeated lack of control can dampen attempts to resist. Dependence matters too: when someone controls your food and safety, you pay close attention to them. Rarely is one factor responsible, and no single theory fits every case.

How common is Stockholm syndrome among hostages?

Less common than people assume. A frequently quoted FBI figure, discussed in a 1999 article by G. Dwayne Fuselier, is about 8 percent of hostage victims, though definitions and data quality vary, so the number should be treated as approximate. Many hostages feel intense fear or anger toward their captors and never develop any bond. The popular idea that bonding is the usual outcome comes largely from a few famous cases and from movies. Police negotiators are trained not to assume the pattern, because assuming it can lead them to misread what a hostage is actually experiencing.

Can Stockholm syndrome happen in abusive relationships?

Sometimes the dynamics are similar, though experts usually prefer the term trauma bonding. Abusive relationships often involve power imbalance, isolation, and cycles of cruelty and kindness, which can create strong attachment. Researchers such as Dee Graham have argued that survival logic seen in hostage cases also operates in battered women, while others caution that the evidence is mostly clinical and that the hostage-based term can be misleading. The important thing is not the label but the safety: coercive control and abuse can be dangerous, and leaving may be the riskiest time, so a safety plan with a trained advocate is wise.

Did Patty Hearst have Stockholm syndrome?

That’s disputed. Patty Hearst was kidnapped in 1974 and later took part in a bank robbery alongside her captors. Her defense argued she acted under coercion and fear, and many commentators invoked Stockholm syndrome to explain her behavior. Prosecutors argued she’d joined voluntarily, and she was convicted in 1976. Her sentence was commuted in 1979 and she was pardoned in 2001. Because the syndrome isn’t a formal diagnosis and no one can read her mind, the question can’t be answered definitively, and the case is better seen as an example of how the label gets used in public debate.

What’s the difference between Stockholm syndrome and trauma bonding?

Stockholm syndrome originally described hostages or captives who felt positive toward captors during a short, life-threatening event. Trauma bonding, a term associated with research by Dutton and Painter on intermittent abuse, describes attachment that develops through repeated cycles of mistreatment and kindness over time, typically in relationships with a power imbalance. They overlap, and people sometimes use them interchangeably, but trauma bonding is generally preferred for ongoing abusive relationships because it comes with more research. Neither is an official diagnosis, and both describe responses to coercive situations, not personal weakness.

How do you help someone who defends their abuser?

Stay calm and keep the relationship intact. Arguments and ultimatums usually trigger defensiveness and can push the person back toward the harmful partner or captor. Listen, validate their feelings without endorsing the abuse, and offer practical help such as transport, a place to stay, or information about advocates. Encourage contact with trained professionals who understand coercive control, and respect that leaving is a process that can take several attempts. If there’s immediate danger, call emergency services. Patience matters; the bond often loosens only when the person feels safe enough to look at it.

Bibliography

• Freud, A. (1936). The ego and the mechanisms of defence. Hogarth Press.
• Fuselier, G. D. (1999). Placing the Stockholm syndrome in perspective. FBI Law Enforcement Bulletin, 68(7), 22-25.
• Graham, D. L. R., Rawlings, E. I., & Rigsby, R. K. (1994). Loving to survive: Sexual terror, men’s violence, and women’s lives. New York University Press.
• Grubb, A. (2026, August 28). Stockholm syndrome isn’t a diagnosis: So why do we still treat it as one? The Conversation.
• Herman, J. L. (1992). Trauma and recovery: The aftermath of violence, from domestic abuse to political terror. Basic Books.
• Namnyak, M., Tufton, N., Szekely, R., Toal, M., Worboys, S., & Sampson, E. L. (2008). “Stockholm syndrome”: Psychiatric diagnosis or urban myth? Acta Psychiatrica Scandinavica, 117(1), 4-11.
• Ochberg, F. M. (1978). The victim of terrorism: Psychiatric considerations. Terrorism, 1(2), 147-168.
• Seligman, M. E. P. (1975). Helplessness: On depression, development, and death. W. H. Freeman.
• Strentz, T. (1980). The Stockholm syndrome: Law enforcement policy and hostage behavior. Annals of the New York Academy of Sciences, 347(1), 137-150.